Spontaneous bacterial peritonitis (SBP) is defined as an intraperitoneal infection of ascitic fluid without underlying anatomic or pathologic cause. The term was initially described by Conn in 1964 who postulated that translocation of enteric pathogens in decompensated cirrhotic patients could lead to peritonitis and bacteremia, a vastly underreported syndrome at the time [1]. Initial reports suggested an exceedingly high mortality rate of greater than 90 % associated with SBP. Despite advances in management and more widespread use of antimicrobial prophylaxis, mortality continues to range from 10 to 30 % making it a significant cause of death in patients with end-stage liver disease [2].

At least 90 % of cases of SBP are monomicrobial with the most common causative organisms including gram negative enteric flora such as Escherichia coli and Klebsiella pneumoniae , as well as various Streptococcal species [3]. Studies seem to suggest a role for bacterial overgrowth and delayed intestinal motility as predisposing factors for the development of SBP. A study by Chang, et al. from 2003 compared cirrhotic patients with and without a history of SBP and found significantly higher rates of bacterial overgrowth and small intestinal dysmotility as assessed by hydrogen breath testing and small bowel manometry respectively among patients with a history of SBP [4]. The study was limited by significantly higher Child-Pugh score in the group with a history of SBP. A subsequent analysis using jejunal aspirates from cirrhotic patients found an association between bacterial overgrowth and acid-suppressive therapy, but no association with the development of SBP [5]. Independent of bacterial overgrowth concerns, the phenomena of bacterial translocation of enteric pathogens in patients with cirrhosis has been well documented. For example, analysis of mesenteric lymph node sampling has revealed increased levels of bacteria in cirrhotic compared to noncirrhotic patients [6].

SBP typically develops in patients with advanced liver disease, and risk is directly proportional to Child-Pugh score which has been shown to be an independent risk factor for bacterial infections in general including SBP [7]. An elevated Model for End-Stage Liver Disease (MELD) score has also been implicated. A retrospective case–control analysis found that for every 1-point increase in MELD score there is an associated 11 % increased risk of developing SBP [8].

There has also been considerable interest in the role of acid-suppressive therapy in the development of SBP. A retrospective case–control analysis of cirrhotic patients admitted with SBP compared to matched control cirrhotic patients admitted for other reasons found on multivariate analysis that PPI use was associated with an increased risk of SBP (odds ratio 4.31, 95 % CI 1.34–11.7) [9]. A more recent large, multicenter prospective analysis found no increase in risk of SBP among patients taking proton-pump inhibitors [10].

Additional known risk factors predisposing to the development of SBP include: low ascitic fluid total protein concentrationcoagulopathy, hyperbilirubinemia, gastrointestinal hemorrhage, and prior episode of SBP [11–14]. Several randomized trials were analyzed and based on results showing mortality benefit and reduced rate of infection, primary prophylaxis with antibiotics to prevent SBP is recommended in patients with cirrhosis and ascites whose ascitic total protein level is less than 1.0 g/dL, in addition to evidence of either impaired renal function (serum creatinine ≥1.2 mg/dL, blood urea nitrogen ≥25 mg/dL, or serum sodium ≤130 mEq/L) or hepatic failure (Child-Pugh score ≥9 or total bilirubin ≥3 mg/dL) [14, 15]. Genetic variability in terms of inflammatory signaling may play a role in terms of SBP risk as well. A recent genotypic analysis showed that specific variants of a Toll-Like Receptor 4 gene were associated with lower serum levels of tumor necrosis factor-alpha (TNF-α) and a significantly decreased risk of severe bacterial infections among patients awaiting liver transplant. Zero deaths as a result of severe bacterial infections were observed while awaiting transplant in this group [16].

Immune dysfunction is thought to play a significant role in cirrhotic patients by predisposing them to the development of infections including SBP. For example, complement deficiency has been well documented in the cirrhotic population and has been demonstrated to correlate with elevated rates of SBP [17]. A variety of other host factors play a role as well including malnutrition, decreased phagocyte activity, neutrophil dysfunction, and altered inflammatory cytokine levels [18].

True SBP in patients with ascites due to noncirrhotic causes , or elevated ascitic protein levels >2.5 g/dL is very uncommon and limited in the literature to case reports and small series [19]. Typically when there is an infection in this setting, an underlying predisposing anatomic defect leading to secondary bacterial peritonitis should be sought.

SBP tends to develop only in cases of clinically apparent and preexisting ascites. Its most common presenting features include abdominal pain, fevers, encephalopathy, diarrhea or ileus, and hemodynamic instability or sepsis, and it should be strongly considered in any patient with cirrhosis and clinically apparent ascites who presents with these complaints. Common laboratory findings include peripheral leukocytosis with or without left shift, acidosis, and renal failure. A significant number of patients present without symptoms, and more routine use of paracentesis for hospitalized cirrhotic patients with ascites has led to increased recognition of this clinical entity. Despite this, mortality rates related to SBP have remained high [20].

SBP is associated with high mortality and early identification and initiation of appropriate antibiotics is critical to mitigating this risk. Based on AASLD guidelines it is recommended that all hospitalized patients with cirrhosis and ascites receive a diagnostic paracentesis during their admission, and paracentesis should also be obtained any time a patient shows clinical signs or laboratory evidence suggestive of possible SBP. Despite these recommendations a recent large retrospective database analysis found that paracentesis remains underutilized, being performed in 61 % of the 17,711 cirrhosis-related hospitalizations included in the study. Patients who received a paracentesis had a 24 % reduction of in-hospital mortality, though their length of stay and cost of hospitalization were slightly higher [21]. Paracentesis should ideally be performed prior to the administration of antibiotics if clinically feasible to prevent false negative results from partially treated sample specimens. Procedural delay should be avoided, and in general, an elevated prothrombin time or international normalized ratio is not a contraindication to diagnostic paracentesis as it has been shown to be safe despite presence of abnormal coagulation factors [22]. Patients with SBP who had paracentesis performed >12 h after admission had a 2.7-fold higher mortality compared to those who received early paracentesis [23].

It is important to utilize sterile technique to prevent contamination from skin flora. Direct bedside inoculation of the sample into culture media has been shown to be superior to conventional inoculation of sample sent to a clinical laboratory with increased sensitivity for diagnosing SBP [24]. Fluid sample should be sent for aerobic and anaerobic culture, cell count with differential, gram stain, and if the initial sampling albumin and total protein levels, in addition to other values relevant on a case by case basis (i.e. cytology for cases of suspected metastatic malignancy, etc.).

The diagnosis of SBP is made by the finding of ≥250 polymorphonuclear (PMN) cells per mm³ with positive culture results and potential causes of secondary peritonitis excluded. Patients meeting criteria based on cell count alone should be presumed to have SBP and treatment with antibiotics should be initiated empirically while awaiting culture results. Correction for grossly blood specimens as occurs during a traumatic paracentesis can be performed by subtracting one PMN from the total count for every 250 red blood cells per mm³ present in the specimen. Additional ascitic fluid chemistries can also be helpful in confirming or excluding the presence of SBP. For example, the serum-ascites albumin gradient (SAAG) is calculated by subtracting the ascitic fluid albumin level from the serum level. A value of >1.1 g/dL indicates the presence of portal hypertension with 97 % certainty. As SBP rarely develops in patients without portal hypertension, a SAAG of <1.1 g/dL makes the diagnosis of SBP unlikely [25]. Additionally, the total protein concentration of ascitic fluid has been shown to inversely correlate with the risk of development of SBP [11].

It is imperative that intravenous antibiotics be initiated promptly following diagnostic paracentesis in cases of suspected SBP due to the high mortality associated with this disease, especially in patients with fevers, abdominal pain, or altered mental status. Patients who are asymptomatic but are found to have bacterascites (defined as the presence of bacteria on culture or gram stain in the setting of PMN count less than 250 cells per mm³) should have a repeat paracentesis performed within 48 h and antibiotics should be initiated if symptoms develop or the PMN count rises above 250 cells per mm³. Initial therapy regimen should consist of broad antimicrobial coverage such as a third generation cephalosporin or floroquinolone, though local resistance patterns should be taken into account when selecting an agent. Choice of antibiotic should be rapidly narrowed when culture results and sensitivities become available to prevent the development of bacterial resistance.

There is a relative dearth of large, prospective, randomized controlled trials available to guide initial antibiotic selection , and a 2009 Cochrane review was unable to provide clear evidence in favor of any specific regimen [26]. One small prospective trial found improved efficacy with IV cefotaxime as compared to ampicillin-tobramycin (clinical cure in 85 % as compared to 56 %, p < 0.02) as well as fewer superinfections and lower rates of renal failure [27]. In general, aminoglycosides are avoided due to their accumulation in the ascitic fluid and thus difficulty in measuring true levels with the potential development of renal failure.

SBP typically responds rapidly to appropriate antibiotic administration , and clinical resolution is typically readily apparent. Typical treatment regimens consist of 5 days of therapy based on findings from a prospective, randomized trial which found equivalent outcomes among patients with SBP who were treated with either 5 or 10 days of antibiotics. Clinical cure was obtained in 93.1 % with 5 days of therapy versus 91.2 % with 10 days of therapy. The groups had similar rates of recurrence (11.6 % and 12.8 % respectively) as well as hospital mortality (32.6 % and 42.5 % respectively), neither of which were statistically significant [28]. Short treatment duration has the added benefit of reduced costs as well as minimizing the development of antimicrobial resistance. After completion of therapy, repeat clinical assessment should be performed and if persistent symptoms or signs of ongoing infection are present (abdominal pain, fevers, leukocytosis, altered mental status) then repeat paracentesis should be performed. Antibiotics should be continued if the polymorphonuclear cell count remains elevated above 250 cells per mm³. For the majority of patients who respond promptly to treatment and have clinical resolution of symptoms, a repeat paracentesis is not typically needed despite previous recommendations of follow-up paracentesis to guide therapy.

Renal failure is common among patient with SBP, occurring in approximately one-third of patients [29]. A recent meta-analysis found that the development of renal failure was the highest independent predictor of mortality among patients with SBP, followed by MELD score. Mortality among those with renal dysfunction in the analysis was 67 % compared to 11 % in patients without renal impairment [30]. Altered renal hemodynamics that occur during infection are thought to play a role, including activation of the renin-angiotensin system as well as release of norepinephrine, effectively reducing renal perfusion [31]. Volume expansion using IV albumin has been shows to improve outcomes among sub-groups of patients with SBP who develop renal failure. A randomized controlled analysis assessed administration of intra-venous albumin concomitantly with antibiotics in patients with SBP who also have evidence of elevated serum creatinine >1.0 mg/dL, blood urea nitrogen >30 mg/dL, or serum bilirubin >4 mg/dL. The treatment arm who received albumin had lower risk of progression to renal failure as well as lower overall mortality [32]. These findings have been corroborated in other randomized trials including a meta-analysis as well [33]. Thus, albumin at a dose of 1.5 g/kg IV within 6 h of diagnosis followed by 1 g/kg IV on day 3 should be administered.

Nonselective beta-blockers are commonly used for prophylaxis of esophageal varices in certain scenarios. This class of medications has myriad influences on systemic hemodynamics and on circulatory reserve. The role of nonselective beta-blocker usage in outcomes among patients with SBP has been investigated. A retrospective analysis of 607 consecutive patients with cirrhosis undergoing paracentesis found that nonselective beta-blocker usage in patients without SBP was associated with increased survival (HR 0.75, 95 % CI 0.581–0.968). Conversely, among patients who were diagnosed with SBP, beta-blocker use reduced survival (HR 1.58, 95 % CI 1.098–2.274), led to more prolonged length of hospitalization (mean 29.6 days per person year versus 23.7 days per person year) and increased incidence of hepatorenal syndrome (24 % versus 11 %) [34]. Therefore, patients who are taking nonselective beta-blockers for prophylaxis of esophageal varices should have this medication discontinued at the time that SBP is first suspected.