Is the acidity of ascitic fluid a reliable index in making the presumptive diagnosis of spontaneous bacterial peritonitis?

Ascitic fluid pH and arterial-ascitic fluid pH gradient were compared to ascitic fluid polymorphonuclear cell count in 84 patients with cirrhotic ascites and in 12 with malignant ascites to assess their role as diagnostic tests for spontaneous bacterial peritonitis and to clarify the relationship between ascitic fluid pH and lactate. Ascitic fluid pH was significantly lower (pH 7.30) in spontaneous bacterial peritonitis (n = 18) and probable spontaneous bacterial peritonitis (n = 12) than in sterile ascites (pH 7.41; n = 54). Since blood pH levels were not different in the presence of infection, arterial-ascitic fluid pH gradient was significantly higher in spontaneous bacterial peritonitis and probable spontaneous bacterial peritonitis than in sterile ascites (0.12 vs. 0.02). The close correlations between arterial-ascitic pH gradient and lactate (r = 0.77, p less than 0.0001), lactate and bicarbonate gradient (r = 0.64, p = 0.003) and arterial-ascitic pH gradient and pCO2 gradient (r = -0.90, p less than 0.0001) suggest that the low ascitic fluid pH may be due to an increase in lactate and CO2. Patients with Escherichia coli-induced spontaneous bacterial peritonitis had significantly lower ascitic fluid pH and higher lactate than those with spontaneous bacterial peritonitis by other organisms. Values of ascitic fluid pH, lactate and arterial-ascitic fluid pH gradient in malignant ascites were similar to those of spontaneous bacterial peritonitis and probable spontaneous bacterial peritonitis. Cutoff points, selected by receiver operating characteristic curves analysis, of 450 per mm3 for polymorphonuclear cells and of 0.07 for arterial-ascitic fluid pH gradient, allow high positive and negative predictive values for spontaneous bacterial peritonitis.(ABSTRACT TRUNCATED AT 250 WORDS)     

Study on ascitic fluid complement 3 level in cirrhotic patients with spontaneous bacterial peritonitis and without spontaneous bacterial peritonitis

BACKGROUND/AIMS: Ascitic fluid Complement 3 (C3) concentration is the most important factor to offer local defense against infection of ascitic fluid. Hepatic synthesis of Complement 3 and its concentration in ascitic fluid is significantly reduced in patients with advanced cirrhosis. The aim of the study was to assess the level of Complement 3 in ascitic fluid in cirrhotic patients with and without spontaneous bacterial peritonitis (SBP) and to identify the group of cirrhotic ascites at risk of developing METHODOLOGY: A prospective case control study was carried out to compare the level of ascitic fluid Complement 3 concentration in patients with SBP (case-group) and without SBP (control-group). Ascitic fluid Complement 3 level was estimated in 15 patients with SBP (case) and another 15 patients without SBP (control). RESULTS: In the study, ascitic fluid Complement 3 concentration was 7.3+/-4.3 mg/dL in patients with SBP and 16.4+/-11.3 mg/dL in patients who did not develop SBP. CONCLUSIONS: Ascitic fluid Complement 3 level is significantly (P=0.009) reduced in cirrhotic patients who develop SBP.     

Outcomes in culture positive and culture negative ascitic fluid infection in patients with viral cirrhosis: cohort study

Background  

Ascitic fluid infection (AFI) in cirrhotic patients has a high morbidity and mortality. It has two variants namely, spontaneous bacterial peritonitis (SBP) and culture negative neutrocytic ascites (CNNA). The aim of this study was to determine the outcome in cirrhotic patients with culture positive (SBP) and culture negative neutrocytic ascites.     

Expression of proinflammatory cytokines and their inhibitors during the course of spontaneous bacterial peritonitis

The aim of this work was the evaluation, in cirrhotic patients with noninfected ascites and with spontaneous bacterial peritonitis (SBP), of serum and ascitic fluid levels of proinflammatory cytokines [interleukin (IL) 1-, tumor necrosis factor (TNF-), and IL6] and antiinflammatory compounds [IL10, soluble IL-1 receptor antagonist (sIL-1Ra), soluble receptors of TNF p55 and p75 (sTNFR55 and sTNFR75), and soluble receptor of IL6 (sIL6R)], as well as their relationship with the outcome of the infection in those with SBP. These molecules were assayed by ELISA in noninfected cirrhotic controls (n = 15), patients with SBP (n = 32), and healthy controls (n = 20). Serum levels of IL6 and of the majority of antiinflammatory mediators, sIL1Ra, sTNFR75, and sIL6R, were higher in control cirrhotic patients compared to healthy subjects. SBP was associated with significantly elevated ascitic fluid levels of every one of the proinflammatory cytokines compared to those in cirrhotic controls. Also, serum levels of IL10 and both TNF receptors and ascitic fluid levels of sIL1Ra and sTNFR55 were higher in patients with SBP compared to cirrhotic controls. Ascitic fluid levels of proinflammatory cytokines decreased rapidly after resolution of the infection; however, nonsignificant changes were detected in ascitic fluid concentrations of antiinflammatory molecules. Thus, elevated levels of antiinflammatory compounds both in noninfected cirrhotic patients and in patients with SBP suggest a regulatory control of the inflammatory process by these molecules in liver cirrhosis patients.     

Prevalence and prognostic significance of bacterascites in cirrhosis with ascites

The prevalence and prognostic significance of bacterascites (BA) were prospectively studied in 443 predominantly HBsAg-positive cirrhotic patients with ascites. Spontaneous bacterial peritonitis (SBP), culture-negative neutrocytic ascites (CNNA), and BA were identified in 12.4%, 8.4%, and 10.8%, respectively. Of these, 67%, 70%, and 71%, respectively, had peritonitis-related signs or symptoms. Among patients with SBP or CNNA, the clinical and laboratory data showed no significant difference between the symptomatic and asymptomatic groups. In contrast, among the patients with BA, the symptomatic group had significantly higher levels of serum total bilirubin and prolonged prothrombin time and significantly lower levels of ascitic fluid total protein than the asymptomatic group. Furthermore, the clinical and laboratory data were relatively similar between patients with asymptomatic BA and those with sterile ascites. In contrast, patients with SBP, CNNA, or symptomatic BA exhibited significantly more severe degrees of liver disease and significantly lower levels of ascitic fluid total protein than those with sterile ascites. There was no statistically significant difference between SBP and bacterascites regarding flora. All patients with SBP, CNNA, or symptomatic BA received antibiotic treatment immediately after paracentesis, as did six of the 14 patients with asymptomatic BA for concurrent respiratory or urinary tract infection, while the remaining eight patients with asymptomatic BA were followed clinically without treatment. Repeated paracentesis in the latter revealed no evidence of SBP or CNNA. The in-hospital mortality for sterile ascites was 22.8%, significantly lower than the 54.5% for SBP, 43.2% for CNNA, and 50% for symptomatic BA, but similar to the 21.4% for asymptomatic BA. In conclusion, 11% of cirrhotic patients with ascites had BA as a complication. Of these, 71% were symptomatic and 29% were asymptomatic. The former might be an SBP variant, while the latter might represent transient colonization of ascitic fluid with bacteria.     

Diagnosing ascites: Value of ascitic fluid total protein, albumin, cholesterol, their ratios, serum-ascites albumin and cholesterol gradient

Abstract Ascitic fluid total protein, albumin, cholesterol, their ascites/serum ratios, serum-ascites albumin and cholesterol gradients were measured for their ability to differentiate cirrhotic, malignant and tuberculous ascites in 76 patients. The mean ± s.d. ascitic fluid total protein, albumin, cholesterol, their respective ascitic fluid/serum ratios in cirrhotic ascites were lower than malignant and tuberculous groups ( P < 0.001 for each). The difference between malignant and tuberculous groups was significant for ascitic fluid/serum total protein ( P < 0.05) and ascitic fluid/serum albumin ( P < 0.01) only. Mean serum-ascites albumin gradient in cirrhotics was higher than in the malignant and tuberculous groups ( P < 0.001 for each). The difference between malignant and tuberculous groups was significant ( P < 0.01). Mean ± s.d. serum-ascites cholesterol gradient in cirrhotics was higher than that in malignant and tuberculous groups ( P < 0.001 for each). The difference between malignant and tuberculous groups was also significant ( P < 0.01). Both serum/ascitic fluid total protein less than 0.5 and ascitic fluid cholesterol less than 55 mg/dL had 94% diagnostic accuracy for differentiating cirrhotic from malignant and tuberculous ascites. Serum ascitic fluid albumin gradient greater than 1.1 g/dL, ascitic fluid/serum albumin less than 0.65 and ascitic fluid albumin less than 2 g/dL had diagnostic accuracy of 92, 92 and 91%, respectively. Ascitic fluid total protein had diagnostic accuracy of 88%. None of the tests was able to differentiate between malignant and tuberculous ascites. Measurement of ascitic fluid cholesterol concentration is a simple method of differentiating cirrhotic from non-cirrhotic ascites.     

Ascitic pH and infection in alcoholic cirrhosis

The pH values of 108 samples of ascitic fluid in 94 alcoholic cirrhotic patients were analyzed in order to assess their diagnostic and prognostic value. The mean pH value of ascitic fluid was significantly lower (p less than 0.001) in patients with spontaneous bacterial peritonitis (7.23 +/- 0.22) or with suspected diagnosis of spontaneous bacterial peritonitis (7.29 +/- 0.15) than in patients with sterile ascites (7.45 +/- 0.06). However, there was an important overlap between these groups. In patients with and without spontaneous bacterial peritonitis, measurement of the difference between blood and ascitic pH was more discriminative than the ascitic pH alone: a difference of 0.10 or more was detected in all patients with spontaneous bacterial peritonitis, in 2 of 5 patients with suspected diagnosis of spontaneous bacterial peritonitis and in 3 of 97 patients with sterile ascites. When the ascitic pH value was lower than 7.15, death occurred rapidly. Ascitic pH rapidly increased when treatment of spontaneous bacterial peritonitis was clinically effective. These results suggest that measurement of pH in ascitic fluid is contributive to the diagnosis and prognosis of spontaneous bacterial peritonitis in alcoholic cirrhosis.     

Leptin levels in the differential diagnosis between benign and malignant ascites

AIM： To evaluate the role of leptin levels in the differentia diagnosis of ascites.
METHODS： Ascitic leptin, TNFα and serum leptin levels were measured in 77 patients with ascites （35 with malignancies, 30 cirrhosis and 12 tuberculosis）. Control serum samples were obtained from 20 healthy subjects. Leptin and TNFα levels were measured by EUSA. Body mass index （BMI） and percentage of body fat （BFM） by skin fold measurement were calculated for all patients and control groups. Peritoneal biopsy, ascites cytology and cultures or biochemical values were used for the diagnosis of patients.
RESULTS： In patients with malignancies, the mean serum and ascites leptin levels and their ratios were significantly decreased compared to the other patient groups and controls. In tuberculosis peritonitis, ascitic fluid TNFα levels were significantly higher than malignant ascites and cirrhotic sterile ascites. BMI and BFM values did not distinguish between patients and controls. CONCLUSION： In patients with malignant ascites, levels of leptin and TNFα were significantly lower than in patients with tuberculous ascites.     

Incidence of spontaneous bacterial peritonitis in patients with ascites. Diagnostic value of white blood cell count and pH measurement in ascitic fluid

During a 21-month period, 65 consecutive patients admitted with ascites were included in a prospective study of the incidence of spontaneous bacterial peritonitis, and paracentesis was performed on admission. The ascitic fluid was cultured, ascitic leucocytes were counted and pH was measured. Bacterial growth was found in five patients with chronic liver disease, who were diagnosed as having spontaneous bacterial peritonitis (SBP), since no intra-abdominal focus could be demonstrated. Thus, the incidence of SBP in this material was 7.7% (95% confidence limits: 2.5-17%). SBP was caused by Escherichia coli (n = 3), coagulase negative staphylococcus (n = 1), and Bacteroides species (n = 1). Abdominal tenderness, abnormal intestinal sounds, fever and hepatic encephalopathy were equally frequent in the group with SBP and in patients with sterile ascites. Infection was not anticipated in any of the patients with SBP. In contrast to several previous studies, neither ascites pH nor ascites leucocyte counts were any help in obtaining a rapid diagnosis. Survival time of patients with SBP was significantly shorter than of patients without SBP.     

Ascitic fluid polymorphonuclear cell count and serum to ascites albumin gradient in the diagnosis of bacterial peritonitis

The analysis of ascitic fluid has been complicated by several recently reported new tests. To simplify this assessment, we evaluated nine parameters prospectively and simultaneously in blood and ascitic fluid from 285 patients with ascites to determine which were the most reliable for immediate diagnosis of the etiology of the ascites and of its complications. Subjects were first divided into four groups: sterile cirrhotic ascites (n = 201), spontaneous bacterial peritonitis (n = 41), malignant ascites (n = 34), and miscellaneous ascites (n = 9). An ascitic fluid polymorphonuclear count greater than 500/microliters was the test with the greatest accuracy (96%) for the diagnosis of spontaneous bacterial peritonitis. Neither the most precise cutoff values for ascitic fluid pH (less than 7.32) and ascitic fluid lactate (greater than 32 mg/dl), nor their respective blood-ascitic fluid gradients (greater than 0.11 and less than -20 mg/dl) were more reliable indexes of spontaneous bacterial peritonitis, mainly due to the decreased ascitic fluid pH and increased ascitic fluid lactate observed in malignant ascites, tuberculous peritonitis, and pancreatic ascites. A blood-ascitic fluid albumin gradient less than 1.1 g/dl was the most accurate parameter for the diagnosis of malignant ascites (diagnostic efficacy, 93%). Therefore, the etiologic analysis of ascitic fluid might be simplified and the single practice of two tests, ascitic fluid polymorphonuclear cell count and blood-ascitic fluid albumin gradient, provides immediately useful information.     

Development of an experimental model of induced bacterial peritonitis in cirrhotic rats with or without ascites

BACKGROUND: Spontaneous bacterial peritonitis (SBP) is a severe complication of cirrhotic patients associated with a high mortality. AIM: To develop an available experimental model of induced bacterial peritonitis in cirrhosis. MATERIAL AND METHODS: Sprague-Dawley rats with carbon-tetrachloride-induced cirrhosis with (N=22) or without (N=101) ascites were randomized to receive an intraperitoneal administration of different concentrations of Escherichia coli (E. coli) diluted in 1 mL of sterile water in ascitic rats and in different volumes in nonascitic rats. A subgroup of nonascitic animals received ceftriaxone 4 h after E. coli inoculation. Mortality of rats was evaluated 24 h after bacterial inoculation. RESULTS: None of the rats receiving sterile water alone and only one infected with 10(7) cfu of E. coli died. Ascitic rats showed a lower mortality rate than nonascitic rats infected with 10(8) or 10(9) cfu of E. coli (P<0.05). Mortality was higher with 10(9) cfu than with 10(8) cfu of E. coli in ascitic (P NS) and nonascitic (P<0.01) rats. A trend was noted to ward higher mortality in nonascitic rats inoculated with 10(8) cfu with increasing water volumes. A marked peritoneal polymorphonuclear cell response was observed 4 h after E. coli injection in both ascitic and nonascitic rats. Antibiotic therapy significantly reduced the mortality rate of rats infected with 10(8) cfu (P<0.01). CONCLUSIONS: This experimental model of induced bacterial peritonitis in cirrhosis with or without ascites may represent a useful tool for the study of pathogenic events postinfection and for the design of new therapeutic strategies to treat patients with SBP.     

Rapid diagnosis of spontaneous bacterial peritonitis by use of reagent strips

We studied the use of reagent strips for diagnosis of spontaneous bacterial peritonitis (SBP) in cirrhotic patients with ascites. A reagent strip for leukocyte esterase designed for the testing of urine with a colorimetric 5-grade scale (0 to 4) was used to evaluate ascitic fluid in 228 nonselected paracentesis performed in 128 cirrhotic patients. We diagnosed 52 SBP and 5 secondary bacterial peritonitis by means of polymorphonuclear cell count and classical criteria. When we considered positive a reagent strip result of 3 or 4, sensitivity was 89% (51 of 57), specificity was 99% (170 of 171), and positive predictive value was 98%. When we considered positive a reagent strip result of 2 or more, sensitivity was 96% (55 of 57), specificity was 89% (152 of 171), and negative predictive value was 99%. In conclusion, the use of reagent strips is a rapid, easy to use, and inexpensive tool for diagnosis of ascitic fluid infection. A positive result should be an indication for empirical antibiotic therapy, and a negative result may be useful as a screening test to exclude SBP.     

Mean platelet volume is a useful indicator of systemic inflammation in cirrhotic patients with ascitic fluid infection

Aim. Ascitic fluid infection (AFI) consists primarily of two variants, namely, culture-negative neutrocytic ascites and spontaneous bacterial peritonitis (SBP). Mean platelet volume (MPV) has begun to be used as a simple and inexpensive indicator of inflammation in some diseases. We aimed to analyse whether platelet size alterations would be useful in predicting AFI in cirrhotic patients.Material and methods. A total of 135 patients with ascites due to cirrhosis and 55 control subjects were enrolled in this study. According to ascitic fluid analysis, 58 patients were considered to have AFI. MPV and inflammatory parameter values were determined for all study participants. The ability of MPV values to predict AFI in cirrhotic patients was analysed using receiver operator characteristic (ROC) curve analysis.Results. A statistically significant increase in MPV levels was observed in cirrhotic patients with AFI compared to cirrhotic patients without AFI and healthy controls (p < 0.001). A statistically significant increase was observed in the AFI group with respect to MPV, C-reactive protein (CRP) and white blood cell (WBC) levels. ROC curve analysis suggested that the optimum MPV level cut-off point for cirrhotic patients with AFI was 8.45, with a sensitivity, specificity, negative predictive value (NPV), and positive predictive value (PPV) of 70.7%, 67.5%, 75.4% and 62.1%, respectively (area under curve: 0.768).Conclusion. Our study shows that MPV is increased in cirrhotic patients with AFI. MPV measurement can considered to be an accurate diagnostic test in predicting AFI, possibly due to an ongoing systemic inflammatory response.     

Increasing frequency of Gram-positive bacteria in spontaneous bacterial peritonitis.

AIM: To evaluate the characteristics and possible recent changes of the microbial causes of spontaneous bacterial peritonitis (SBP) in cirrhotic patients. METHODS: We retrospectively evaluated 42 cirrhotic patients with positive ascitic fluid culture and without evidence of secondary peritonitis who were admitted consecutively to our Department between 1998 and 2002. RESULTS: Twenty (48%) of 42 patients with positive ascitic fluid culture were diagnosed during 1998-1999 (period A) and the remaining 22 (52%) patients during 2000-2002 (period B). Gram-negative bacteria were the cause of SBP in 15 (75%) of the 20 patients during period A and in only nine (41%) of the 22 patients during period B (P=0.026). SBP patients with Gram-positive bacteria compared with those with Gram-negative bacteria were less frequently in Child class C (P=0.058) and had significantly higher ascitic fluid protein (P=0.014) and albumin concentrations (P=0.009) and lower ascitic fluid neutrophil count (P=0.008). Resistance to quinolones was detected significantly more frequently in the isolated Gram-positive than Gram-negative bacteria (P<0.001). CONCLUSION: Culture-positive SBP in cirrhotic patients are caused more frequently by Gram-positive bacteria during the recent years, which are, in their vast majority, resistant to quinolones.     

Ascitic fluid infection in patients with hepatitis B virus-related liver cirrhosis: Culture-negative neutrocytic ascites versus spontaneous bacterial peritonitis

Background and Aim:  Ascitic fluid infection (AFI) consists of culture-negative neutrocytic ascites (CNNA) and spontaneous bacterial peritonitis (SBP). The present study compared the clinical characteristics and prognosis of CNNA and SBP in hepatitis B virus (HBV)-related cirrhotic patients.
Methods:  We analyzed 130 consecutive patients hospitalized due to the first episode of AFI between January 1998 and December 2007.
Results:  The mean age of the patients was 52.3 years (88 men, 42 women). Ninety-three patients (71.5%) had CNNA and 37 patients (28.5%) had SBP; 117 patients (90.0%) died after a median survival period of 6.4 months. Patients with CNNA and SBP survived for a median period of 6.9 months and 5.4 months, respectively ( P  = 0.417). Patients with SBP showed higher in-hospital mortality than those with CNNA (16.2 vs 4.3%; P  = 0.031). Binary logistic regression analysis showed that culture positivity of ascitic fluid (CNNA vs SBP) was the only independent predictor of in-hospital mortality ( P  = 0.042). In a Cox regression model for the 120 patients (92.3%) who survived the first episode of AFI, only the Child–Pugh score remained significant for survival ( P  = 0.007), whereas no association was observed for culture positivity of ascitic fluid (CNNA vs SBP) during the first episode of AFI ( P  = 0.752).
Conclusions:  Although in-hospital mortality was higher in patients with SBP than CNNA, the clinical course of the two groups was similar after the first episode of AFI. Thus, liver transplantation should be considered, irrespective of culture positivity of ascitic fluid.     

IL-10、IL-18和内毒素在肝硬化患者自发性细菌性腹膜炎中的作用

目的:探讨肝硬化合并自发性细菌性腹膜炎(spontaneous bacterial peritonitis,SBP)患者血清和腹水中IL-10,IL-18和内毒素水平及意义.方法:采用ELISA方法检测无菌性腹水(sterile ascites,SA)组(32例)和SBP组(45例)血清和腹水IL-10,IL-18水平...     

Analysis of monocyte chemotactic protein-1 gene polymorphism in patients with spontaneous bacterial peritonitis

AIM: To investigate a genetic polymorphism of the monocyte chemotactic protein-1 ( MCP-1) gene in patients with spontaneous bacterial peritonitis (SBP).METHODS: MCP-1 genotyping was performed in 23 patients with SBP and 83 cirrhotic control patients with non-infected ascites.RESULTS: The frequency of carriers of the G-allele was lower in SBP patients but this difference did not reach statistical significance. However, in the subgroup of patients with alcoholic cirrhosis ( n = 80), carriersof the G-allele were significantly less frequent in SBPpatients(38.1%) than in cirrhotic controls (67.8%, P =0.021).CONCLUSION: In patients with alcoholic liver cirrhosis,the -2518 MCP-1 genotype AA is a risk factor for the development of SBP.     

Detection of ascitic fluid infections in patients with liver cirrhosis and ascites

Background and study aimsAscitic fluid infections (AFIs) are the frequent complications of advanced liver disease. Bacterial translocation is considered a key step in the pathogenesis of gut-derived bacterial infections; mainly spontaneous bacterial peritonitis (SBP) in cirrhotic patients. Bacterial DNA (bactDNA) in ascitic fluid and serum has been suggested as a surrogate marker for bacterial translocation. We attempted at the isolation and identification of bacteria in ascitic fluid in cirrhotic patients and the assessment of polymerase chain reaction (PCR) in ascitic fluid and serum.Patients and methodsFifty cirrhotic patients having ascites with no signs of infection were included. Ascitic fluid cultures were obtained from patients. Ascitic fluid and serum were subjected to DNA extraction and PCR for the universal amplification of a region of the 16S ribosomal RNA (16S rRNA) gene to detect bactDNA.ResultsBacteria were isolated from 9 (18%) of the ascitic fluid samples, and were mainly Gram-positive bacteria. BactDNA was detected simultaneously in the ascitic fluid and serum of 17 (34%) patients and in the ascitic fluid of only 2 patients. In a single patient with positive ascitic fluid culture no bactDNA was detected in ascitic fluid or serum. By considering AFIs as a positive ascitic fluid culture and/or the presence of bactDNA in the ascitic fluid and/or serum, ascitic fluid culture could detect 9 out of 20 patients with AFIs (45%), PCR of ascitic fluid could detect 19 out of 20 (95%) while PCR of serum could detect 17 out of 20 (85%). In 10 patients with culture negative non-neutrocytic ascites (CNNNA) bactDNA could be detected in serum and ascitic fluid.ConclusionAFI can be caused by Gram positive as well as Gram negative organisms. A substantial percentage of cases with CNNNA show bactDNA in serum and ascitic fluid. PCR of ascitic fluid should, therefore, be used in the diagnostic workup of suspected cases of ascitic fluid infections.     

Can the protein concentration of the ascitic fluid in ascites predict the occurrence of an infection?

In cirrhotic patients, spontaneous bacterial peritonitis is frequent and severe. This study was performed to determine if low protein concentration in ascitic fluid on admission could predict the occurrence of spontaneous bacterial peritonitis during hospitalization. Ninety-two cirrhotic patients with ascites, without spontaneous bacterial peritonitis were studied. Bacteriologic study and cultures of ascitic fluid were performed on admission and repeated every 5 days, and if any suspicion of infection occurred; 11 patients developed spontaneous bacterial peritonitis during hospitalization. Among the 92 patients in the study, protein concentration in ascitic fluid was initially less than 10 g/l in 45 and 10 of these 45 patients (22 p. 100) developed spontaneous bacterial peritonitis during hospitalization; protein concentration in ascitic fluid was initially greater than 10 g/l in 47 patients; only one of these 47 patients (2.1 p. 100) developed spontaneous bacterial peritonitis during hospitalization. This difference (22 p. 100 vs 2.1 p. 100) was significant (p less than 0.01). Ascitic fluid protein concentration (6.9 +/- 2.3 g/l) was significantly lower (p less than 0.01) in the spontaneous bacterial peritonitis group than in patients without peritonitis (13.8 +/- 10.5 g/l). These results suggest that: 1) ascitic fluid protein concentration on admission is lower in patients who will develop spontaneous bacterial peritonitis during hospitalization than in patients without infection and 2) patients with ascitic fluid protein concentration under 10 g/l on admission represent an high risk group for spontaneous bacterial peritonitis.