Review Article

Spontaneous Bacterial Peritonitis – Bacteriology, Diagnosis, Treatment, Risk Factors and Prevention

J. B. Dever; M. Y. Sheikh


Aliment Pharmacol Ther. 2015;41(11):1116-1131. 

In This Article


Clinical Perspective

The diagnosis of SBP is achieved in the absence of any apparent aetiology of infection or secondary peritonitis as may be seen with gastrointestinal tract perforation, appendicitis, diverticulitis or cholecystitis. SBP is not a clinical diagnosis, and it cannot be made without ascitic fluid analysis. Interpreting ascitic fluid results namely cell count and differential to calculate the polymorphonuclear (PMN) count, which is the result of multiplying the total ascitic fluid white blood cell count by the neutrophil count, is paramount. A PMN count greater than 250 cells per mm3 provides a preliminary diagnosis, and although culture results are often negative, historically, a positive ascitic fluid culture confirms the diagnosis of SBP.[43] Variants to this classic definition include culture-negative neutrocytic ascites (CNNA), which was originally defined as an ascitic PMN count greater than 500/mm3 in the setting of a negative culture; however, now keeps with the 250/mm3 threshold and is a commonly encountered phenotype of SBP. This phenomenon may represent resolution of infection, small amounts of bacteria or its by-products in ascites or poor culture processing.[44] Despite sound technique with performing paracentesis, ascitic fluid culture is negative in at least 40% of cases with an elevated PMN count.[45] CNNA is a true infection and should be treated the same as SBP.

Spontaneous bacterial peritonitis is the most common bacterial infection in hospitalised patients with cirrhosis and ascites who are not receiving antibiotic prophylaxis.[46,47] Acquisition of SBP occurs in various clinical settings which have different clinical implications (Table 3). The diagnosis should be considered in any patient with cirrhosis, ascites and clinical deterioration and although rare, can be clinically silent without apparent signs or symptoms. The prevalence of SBP in 427 cirrhotic asymptomatic out-patients, as defined by neutrocytic ascites (absolute neutrophil count >250 cells/mm3), was 3.5%.[29] Conversely, there should be a high level of suspicion for infection when evaluating patients with Child–Pugh classification C liver disease with fever, abdominal pain and/or confusion.[43,48] Diagnostic paracentesis should be performed without delay, ideally within 6 hours of patient evaluation, and before the use of antibiotics. Early and judicious use of antibiotics can be life saving especially for patients with end-stage liver disease defined by one or more of the following: history of ascites, variceal haemorrhage, hepatic encephalopathy, hepatopulmonary syndrome, hepatorenal syndrome, SBP or hepatocellular carcinoma.

Laboratory Diagnosis

The current gold standard for cell counting as per the College of American Pathologists is by light microscopy using a manual counting chamber, which is operator dependent and can predispose to inconsistent results among laboratory technicians with varying expertise, yet has proven to be more reliable than automated cell counting. Automated cell counting for PMNs was shown to be accurate, reliable and faster compared to manual counting;[49] however, this is not a universal experience and current guidelines do not recommend this modality. At particularly low cell counts, around 250, automated cell counters have been prone to false-positive results.[50] Flow cytometry is a newer form of technology and has shown to be a suitable alternative for rapid ascitic fluid cell counting with a sensitivity and specificity of 100% for detecting PMN count greater than 250 cells/mm3.[51]

Urinary Reagent Strips

Urinary reagent strips were proposed as a quick method to diagnose SBP,[52,53] but a large multicentre study revealed that the Multistix 8SG strip is an unacceptable test because of poor sensitivity, positive predictive value and inability to rule out infection.[54] Ascitic fluid lactoferrin (AFLAC) has also been supported as an initial screening test. Lactoferrin is a product and marker of PMN activity and when detected in ascitic fluid has shown high sensitivity (95.5%) and specificity (97%) for SBP,[55] however timing of quantitative measurements using an enzyme-linked immunosorbent assay have not been reported. Its diagnostic utility is questioned and manual cytological examination remains the standard method for determining PMN concentration.[56]


Ascitic fluid is obtained by transabdominal needle puncture at one of several locations, the avascular infraumbilical midline, right lower quadrant or preferably the left lower quadrant,[57] and its analysis is one of the most important steps in the evaluation of patients suspected of having SBP. Paracentesis should be performed in all patients with cirrhosis and ascites who require emergency room care or hospitalisation, report symptoms, or demonstrate signs of clinical deterioration such as confusion or gastrointestinal bleeding.[58] Although paracentesis is vital in the evaluation of such patients and has been associated with nonsignificant increased short-term survival, it is underutilised, performed in only 60% of hospitalised patients with cirrhosis and ascites.[59] In patients who do undergo paracentesis, delay in procuring ascitic fluid for analysis has led to a 2.7-fold increased risk of in-hospital mortality in patients with SBP.[60] The clinical usefulness of paracentesis is less important in the out-patient setting as the diagnostic yield for SBP or its variants is less than 4%.[61]

Paracentesis has consistently been shown to be safe without using ultrasound (US) and in the setting of an elevated prothrombin time[62] and remains the most cost-effective method of diagnosis.[63] Major complications such as bleeding and infection have been reported to be higher in other prospective series, occurring at 1.6%, which resulted in two deaths out of 515 paracenteses.[64] Using US during parcentensis is practised by house staff at some medical centres such as the University of California at San Francisco, and support for its use has likely derived from recent changes in standards for other procedures requiring needle puncture such as central venous line catheter placement, and case studies with poor outcomes.[65] Evidence to recommend US for paracentesis is limited. A retrospective study supported by Sonosite, demonstrated the risk of bleeding from paracentesis was decreased from 1.25% to 0.27% with the use of US out of 70 000 paracentesis patients.[66]

Correcting coagulopathy prior to paracentesis often using arbitrary INR thresholds of 2–3 is not based on supporting evidence but is commonly employed in clinical practice to avoid haemorrhagic complications. Liberal or unnecessary use of fresh frozen plasma may, in fact, lead to complications seen with blood product administration. Paracentesis was safe in 1100 patients with an INR as high as 8.7 and platelet count as low as 19 000.[67] Patients with disseminated intravascular coagulation (DIC) generally pose a greater risk for complications and should be managed with greater diligence.

Paracentesis should be performed by well-trained personnel, not limited to physicians, who have completed 3–10 paracentesis under supervision by an experienced clinician.[67] Caution is quintessential for performing procedures and should be employed especially in patients with ileus, prior abdominal surgery or bleeding diatheses. In such cases, US guidance might be particularly helpful. Ascitic fluid tests should include cell count with differential, Gram stain, culture, total protein and albumin to determine the serum-ascites albumin gradient (SAAG) if not already known.

Proper handling of ascitic fluid requires bedside inoculation of at least 10 mL into aerobic and anaerobic blood culture bottles, which increases the yield of positive culture results by 91% and speeds time of detection.[68] For cell count, approximately 1 mL of fluid should be injected into a purple-top ethylenediaminetetraacetic acid (EDTA), which is typically used for determining serum complete blood cell count and often included in paracentesis kits, to avoid clotting and inaccurate interpretation. Although ascitic fluid appearance such as haziness had a sensitivity of 98% in detecting SBP and clear ascitic fluid was less likely to be infected based on a retrospective review of more than 900 samples,[69] its appearance cannot be used as a reliable surrogate marker to achieve or exclude the diagnosis of SBP.[70] Furthermore, low clinical suspicion for SBP does not preclude the necessity for paracentesis. Obtaining a positive ascitic fluid culture result is highly variable ranging from 3.5% in the emergency room[69] to 60% in hospitalised patients,[45] however if performed properly prior to antibiotic administration, cultures should yield bacterial growth in more patients with SBP.