Spontaneous bacterial peritonitis — from clinical suspicion through diagnostic tap, empiric therapy, albumin, and prophylaxis
Spontaneous bacterial peritonitis (SBP) is an infection of ascitic fluid without any identifiable intra-abdominal surgical source, arising from bacterial translocation across a permeable gut wall in the setting of cirrhotic portal hypertension. Because its presentation is frequently subtle or entirely absent, and because untreated SBP carries substantial short-term mortality, guidelines mandate an aggressively low threshold for diagnostic paracentesis in any cirrhotic patient with ascites.
SBP develops through a well-characterized sequence tied to the physiology of advanced cirrhosis: portal hypertension causes intestinal congestion and increased mucosal permeability, while cirrhosis-associated immune dysfunction (reduced reticuloendothelial/Kupffer cell clearance, decreased complement and opsonic activity in ascitic fluid) impairs normal bacterial clearance. Enteric bacteria — predominantly gram-negative organisms of gut origin, most commonly Escherichia coli, followed by Klebsiella pneumoniae and gram-positive organisms such as Streptococcus species — translocate across the intestinal wall into mesenteric lymph nodes and eventually seed the ascitic fluid, where the deficient local and systemic immune defenses fail to contain the resulting infection.
This mechanism — translocation from the gut lumen, not a perforated viscus or other surgical source — is what defines SBP as "spontaneous" and distinguishes it from secondary bacterial peritonitis (infection from a perforated viscus, abscess, or other surgical intra-abdominal source), a distinction with major treatment implications discussed further in Stage 3.
Classic presenting features include fever, diffuse abdominal pain and tenderness, and rebound/guarding, but the clinical reality is that SBP's presentation is often incomplete or entirely absent:
• Fever is present in only about half of cases • Abdominal pain/tenderness may be mild, diffuse, or absent altogether given the large volume of ascitic fluid that can mask peritoneal signs on exam • New or worsening hepatic encephalopathy — sometimes the only presenting sign — reflects systemic inflammatory/infectious insult unmasking or worsening underlying hepatic dysfunction • Unexplained acute kidney injury or worsening renal function is a common and easily overlooked presentation, since SBP is a major precipitant of hepatorenal syndrome-type physiology • Approximately 13-30% of patients with culture/PMN-confirmed SBP are asymptomatic at the time of diagnosis, identified only because a diagnostic tap was performed as routine screening
Because of this unpredictable and often subtle presentation, guidelines recommend diagnostic paracentesis be performed liberally: at every hospital admission for a cirrhotic patient with ascites (regardless of admission reason), and for any new or worsening symptom that could plausibly reflect infection — fever, abdominal pain, GI bleeding, new/worsening encephalopathy, or unexplained renal function decline — rather than reserving the tap only for patients with a classic, fully developed clinical picture.
The single most important practical lesson in SBP is that absence of classic peritoneal signs does not exclude the diagnosis. A cirrhotic patient admitted with encephalopathy alone, or with an unexplained creatinine rise, has met sufficient threshold for a diagnostic tap — waiting for fever and rebound tenderness to appear risks a preventable delay in life-saving antibiotics.
The diagnosis of SBP is made — and treatment is initiated — based on the ascitic fluid absolute polymorphonuclear neutrophil (PMN) count, not on a positive culture. A PMN count of 250 cells/mm3 or greater is diagnostic of SBP regardless of what the culture ultimately shows, a deliberately conservative threshold chosen because culture sensitivity is imperfect and treatment cannot safely wait for culture results to return.
Ascitic fluid culture, even when optimally performed (see Stage 3 on bedside blood-culture-bottle inoculation), remains falsely negative in a substantial proportion of true SBP cases — reported sensitivity in the range of 50-65% — because the bacterial burden in ascitic fluid is often quite low (frequently under 1 organism per mL), well below the reliable detection threshold of standard culture techniques applied to a modest fluid volume. Waiting for a positive culture before starting antibiotics would therefore miss a large fraction of genuine infections and delay treatment in a condition where prompt antibiotic initiation measurably improves survival.
Because of this, the ascitic fluid absolute PMN count has been adopted as the primary diagnostic and treatment-triggering criterion: a PMN count ≥250 cells/mm3 is diagnostic of SBP and mandates immediate empiric antibiotic therapy, regardless of whether culture ultimately grows an organism. This threshold was chosen (rather than a lower or higher cutoff) based on studies correlating PMN counts with clinical outcomes and treatment response, balancing sensitivity for true infection against the risk of over-treating sterile ascites.
When PMN ≥250 but culture is negative, this is termed culture-negative neutrocytic ascites (CNNA) — it is managed identically to culture-positive SBP, with a full course of empiric antibiotics, because the elevated PMN count itself is taken as sufficient evidence of infection. Conversely, "bacterascites" (positive culture but PMN <250) represents a different, usually transient or self-resolving colonization state and is managed with repeat paracentesis and antibiotics only if the patient becomes symptomatic or the PMN count rises on repeat testing, rather than automatic treatment of every positive culture in isolation.
The serum-ascites albumin gradient (SAAG), calculated as serum albumin minus ascitic fluid albumin drawn on the same day, is used to confirm that ascitic fluid is due to portal hypertension (SAAG ≥1.1 g/dL, with >97% accuracy for portal hypertensive causes including cirrhosis) — supporting a cirrhotic/SBP-consistent context — versus non-portal-hypertensive causes of ascites (SAAG <1.1 g/dL, seen in peritoneal carcinomatosis, tuberculous peritonitis, pancreatic ascites, and other etiologies) which would prompt a different diagnostic pathway entirely.
Distinguishing primary SBP from secondary bacterial peritonitis (infection from a perforated viscus, intra-abdominal abscess, or other surgical source) is a critical branch point, because secondary peritonitis requires surgical source control in addition to antibiotics — antibiotics alone will fail. Clues suggestive of a secondary process rather than spontaneous SBP include: markedly elevated ascitic fluid total protein (>1 g/dL), very high LDH (exceeding serum LDH), very low ascitic glucose (<50 mg/dL), and — most specifically — a polymicrobial culture result (SBP is almost always monomicrobial; growth of multiple organisms, particularly including anaerobes or fungi, strongly suggests a perforated viscus). When these features are present, or when a patient fails to improve on appropriate antibiotics as expected, urgent cross-sectional imaging to evaluate for a surgical source is warranted rather than simply escalating or changing antibiotics.
Once the ascitic fluid PMN count meets the diagnostic threshold, empiric antibiotic therapy is started immediately — it should not wait for culture confirmation. Third-generation cephalosporins (cefotaxime or ceftriaxone) are the first-line empiric choice, reflecting both their reliable coverage of the typical causative organisms and, critically, their favorable renal safety profile in a patient population where nephrotoxic agents carry outsized risk.
Ascitic fluid culture yield is significantly improved by inoculating fluid directly into standard aerobic and anaerobic blood culture bottles at the bedside at the time of paracentesis, rather than sending fluid to the laboratory in a plain sterile container for later plating. This technique — analogous to blood culture collection — increases the volume-to-broth ratio and provides an immediate growth-promoting medium, and has been shown in multiple studies to substantially improve culture sensitivity (from roughly 50% with conventional plating methods to closer to 80% with bedside bottle inoculation) even though, as discussed in Stage 2, culture remains an imperfect and non-time-critical adjunct to the PMN-based diagnosis and treatment decision.
Approximately 10 mL of ascitic fluid is typically inoculated into each of the aerobic and anaerobic bottles at the bedside immediately upon collection, in addition to fluid sent for cell count/differential, albumin, total protein, and glucose.
Third-generation cephalosporins — cefotaxime (typically 2 grams IV every 8 hours) or ceftriaxone (2 grams IV once daily) — are the standard first-line empiric choice for community-acquired SBP, chosen for their reliable spectrum against the predominant causative organisms (E. coli, Klebsiella pneumoniae, and streptococcal species account for the large majority of isolates) and their well-established efficacy in the foundational clinical trials establishing modern SBP management. A typical treatment course is 5 days, shown in randomized trials to be non-inferior to longer courses (historically 10 days) for uncomplicated SBP with adequate clinical response, allowing for a shorter, less resource-intensive treatment duration in appropriately responding patients.
In patients with prior fluoroquinolone prophylaxis exposure, healthcare-associated or nosocomial SBP, or in regions/institutions with higher rates of multidrug-resistant organisms, broader empiric coverage (e.g., piperacillin-tazobactam or a carbapenem) may be warranted, guided by local antibiogram data and individual risk factors — reflecting the same antimicrobial stewardship principles applied to other serious infections in the era of rising resistance.
Aminoglycosides (gentamicin, tobramycin) are specifically avoided in the empiric treatment of SBP despite their activity against gram-negative organisms, because cirrhotic patients — who already have impaired renal autoregulation and are frequently volume-depleted or hemodynamically labile — are at substantially elevated risk of aminoglycoside-induced nephrotoxicity, which can precipitate or accelerate hepatorenal syndrome-type renal failure. This risk-benefit calculation, specific to the cirrhotic population, is one of the clearest examples in hepatology of a drug class being avoided not for lack of antimicrobial efficacy but because of disproportionate harm in this particular patient population.
Cefotaxime/ceftriaxone should be started as soon as the diagnostic PMN threshold is met — not after culture results return. Every hour of delay in a patient with true SBP measurably worsens outcomes, and the combination of prompt empiric third-generation cephalosporin plus judicious avoidance of nephrotoxic alternatives defines the modern standard of care.
SBP is one of the most important precipitants of hepatorenal syndrome-type acute kidney injury in cirrhosis, and this renal complication — not the infection itself — is the dominant driver of mortality in SBP. The landmark Sort et al. trial (NEJM, 1999) demonstrated that adding IV albumin to antibiotic therapy, in a specific two-dose regimen, significantly reduces the incidence of renal impairment and improves survival, particularly in patients at highest risk.
Bacterial infection in cirrhosis triggers a marked exaggeration of the already-present splanchnic arterial vasodilation and systemic inflammatory response, further reducing effective circulating arterial volume and activating intense compensatory vasoconstriction of the renal circulation — the same underlying mechanism as hepatorenal syndrome, here precipitated acutely by infection rather than arising spontaneously. This renal hypoperfusion, if not corrected, can progress from a functional, potentially reversible state to established acute tubular injury and renal failure, which historically has been the dominant driver of mortality in SBP — patients who died from SBP typically died from this renal complication and its downstream multi-organ consequences, not directly from the infection itself.
This mechanistic understanding — that renal circulatory compromise, not the infection per se, drives most SBP mortality — is what motivated investigation of a volume-expansion strategy (analogous to the albumin rationale in post-paracentesis circulatory dysfunction, though a distinct clinical scenario) as an adjunct to antibiotic therapy.
Sort and colleagues (New England Journal of Medicine, 1999) randomized 126 patients with cirrhosis and SBP to receive either antibiotic therapy alone (cefotaxime) or cefotaxime plus IV albumin, dosed at 1.5 g/kg of body weight within 6 hours of diagnosis (day 1) and 1 g/kg on day 3. The albumin group had a significantly lower incidence of renal impairment (approximately 10% versus approximately 33% in the antibiotic-alone group) and significantly lower in-hospital mortality and 3-month mortality — a result that has been reproduced and extended in subsequent studies and meta-analyses, establishing albumin as a standard adjunct in SBP management for appropriately selected patients.
Subsequent work has refined patient selection, since the mortality/renal benefit of albumin appears concentrated in patients at highest baseline risk of renal complications, rather than being uniformly beneficial across all SBP patients regardless of severity. Current practice, reflecting this refinement, restricts routine albumin administration to patients meeting at least one of the following criteria at diagnosis: serum creatinine >1 mg/dL, blood urea nitrogen >30 mg/dL, or total bilirubin >4 mg/dL — patients without any of these risk markers derive comparatively little incremental benefit from albumin and may reasonably be treated with antibiotics alone, sparing them the cost and volume burden of albumin infusion. The dosing regimen itself (1.5 g/kg day 1, 1 g/kg day 3) remains the standard when albumin is indicated, mirroring the original trial protocol.
Albumin in SBP is not a universal add-on — it is a targeted intervention for patients with baseline renal or hepatic risk markers (creatinine, BUN, bilirubin) at the time of diagnosis. Applying the Sort et al. criteria correctly identifies exactly the subgroup shown to benefit, avoiding unnecessary albumin exposure in lower-risk patients.
SBP has a strikingly high recurrence rate once a patient has survived a first episode, making secondary antibiotic prophylaxis a near-universal recommendation after any treated episode. A separate, more selectively applied strategy — primary prophylaxis — targets patients who have never had SBP but whose ascitic fluid characteristics and clinical risk factors identify them as being at particularly high imminent risk of a first episode.
Patients who survive an episode of SBP face a markedly elevated risk of a subsequent episode — reported 1-year recurrence rates of approximately 70% in patients who do not receive prophylactic antibiotics after their index episode, reflecting the fact that the underlying predisposing physiology (portal hypertension, gut bacterial translocation, impaired local ascitic fluid immune defenses) persists unchanged after the acute infection resolves.
Daily oral fluoroquinolone prophylaxis — norfloxacin 400mg daily or ciprofloxacin 500mg daily (norfloxacin is not available in all markets, making ciprofloxacin the more commonly used agent where norfloxacin is unavailable) — reduces this 1-year recurrence risk from approximately 70% down to approximately 20%, a substantial and consistently reproduced benefit across multiple trials. Because of this dramatic risk reduction, secondary prophylaxis is recommended indefinitely (or until the ascites resolves, such as after successful liver transplantation) for every patient who has survived an episode of SBP, essentially without exception, making it one of the more universally applied prophylactic antibiotic strategies in hepatology.
Primary prophylaxis — antibiotic prevention in patients who have never had an episode of SBP — is applied more selectively than secondary prophylaxis, reserved for patients identified as being at meaningfully elevated risk of a first episode based on ascitic fluid characteristics and additional clinical risk markers, rather than being given to every cirrhotic patient with ascites:
The principal criterion is a low ascitic fluid total protein (<1.5 g/dL) — low-protein ascites has reduced opsonic and antibacterial activity (lower complement and immunoglobulin content), making it a substantially more permissive environment for bacterial growth once translocation occurs, and low ascitic protein is a well-validated independent predictor of first-episode SBP risk.
Current guidance restricts primary prophylaxis to patients with low ascitic protein (<1.5 g/dL) who also have at least one additional risk factor, reflecting that low protein alone identifies too broad a population to justify indefinite antibiotic exposure (with its attendant costs, resistance-promotion concerns, and risk of selecting for more resistant/difficult-to-treat organisms should breakthrough infection occur). Additional qualifying risk factors typically include: impaired renal function (creatinine ≥1.2 mg/dL, BUN ≥25 mg/dL, or sodium ≤130 mmol/L), advanced liver dysfunction (Child-Pugh score ≥9 with bilirubin ≥3 mg/dL), or a prior episode of SBP being treated as an indication for secondary rather than primary prophylaxis. Patients admitted with acute GI variceal bleeding are managed under a separate, well-established short-course prophylaxis protocol (typically ceftriaxone or a fluoroquinolone for up to 7 days) regardless of ascitic protein level, given the strongly elevated infection risk specifically associated with the bleeding episode itself.
The prophylaxis strategy mirrors the diagnostic philosophy running through the entire SBP pathway: apply the intervention precisely where the evidence shows benefit — universally after a treated episode (secondary prophylaxis), and selectively before a first episode only in the subgroup whose ascitic fluid and clinical risk factors identify genuinely elevated risk (primary prophylaxis) — rather than either under-treating high-risk patients or over-exposing low-risk patients to indefinite antibiotics.