Spontaneous bacterial peritonitis (SBP) is a serious problem in patients with cirrhotic ascites, with an incidence of 10% to 25% on hospital admission (1) and accounting for 5% to 30% of all infections in cirrhotic patients (2, 3), 19% of documented bacteremic episodes (4) and 60% to 75% of all "serious" infections (5, 6). Recent studies (1,7,8) have focused attention on culture techniques, diagnosis, differentiation from secondary peritonitis, pathogenesis and predisposing factors. However, data on prevention and therapy have been scarce.
Gin& et al. in this issue of HEPATOLOGY report the first prospective trial of prevention of SBP by use of longterm antibiotic prophylaxis (9). Is this rational with our present knowledge of the pathogenesis and therapy of SBP?
Most organisms cultured from ascites in patients with SBP (1) are part of the normal aerobic flora of the gut (10-12). Low concentrations (one organism per milliliter) of a single organism are typical (13). SBP with multiple organisms or anaerobes is distinctly uncommon and probably represents secondary bacterial peritonitis. Sixty percent to 80% of the organisms in patients with SBP are aerobic gram-negative bacteria (1, 2, 5, 6, 11, 14-17). Escherichia coli is associated with 40% to 50% of cases of SBP (1, 2, 5, 6, 14), perhaps reflecting the abundance in stool (9-11). However, more frequent extraascitic dissemination with E. coli is compatible with greater virulence (9, 18, 19). The remaining organisms causing SBP are gram-positive cocci; half of which are found in the gut (1). Thus the gut appears to be the repository of most organisms causing SBP.
Multiple mechanisms potentially predispose to SBP. Studies of the microbial flora in cirrhotic patients show no significant qualitative or quantitative difference from normal individuals (20-30). Thus an unusual spectrum or quantity of aerobic organisms or decreased colonization resistance by anaerobic bacteria does not appear to play a role. The decreased serum bactericidal and opsonic activity and abnormal polymorphonuclear cell function demonstrated in cirrhotic patients (31-34) have not been associated with SBP, and the degree of defect appears insufficient for such a role. Intrahepatic and extrahepatic shunting may predispose to SBP by increasing systemic bacteremia from the gut owing to