- Email: [email protected]
ALERT
Copyright © European Association for the Study of the Liver 1997
Journal of Hepatology 1997; 26:839-844 Printed in Denmark • All rights reserved Munksgaard. Copenhagen
Journal of Hepatology ISSN 0168-8278
Early microbiologic diagnosis of spontaneous bacterial peritonitis with BacT/ALEgF Jordi Ortiz, Germfin Soriano, Pere Coil 1, Maria Teresa Novella, Roser Pericas 1, Miriam S/tbat, Ferran Sfinchez 1, Carlos Guarner, Guillem Prats I and Francisco Vilardell Departments o f Gastroenterology and IMicrobiology, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain
Background~Aims: Inoculation of ascitic fluid into conventional blood culture bottles is more sensitive than conventional culture in the diagnosis of spontaneous bacterial peritonitis. BacT/ALERT is an automated colorimetric microbial detection system that has been shown to be faster than conventional blood culture bottles in the diagnosis of bacteremia. The aim of the study was to compare the BacT/ ALERT system with the conventional culture and the conventional blood culture bottles method in the diagnosis of spontaneous bacterial peritonitis. Methods: All the ascitic fluid samples from patients with cirrhosis hospitalized in our Department between September 1992 and May 1994 (n=1032) were prospectively evaluated. In all cases, an aliquot of ascitic fluid was sent for Gram's stain and conventional culture, and 20 ml were inoculated at the bedside into blood culture bottles: 10 ml into conventional blood culture bottles and 10 mi into BacT/ALERT. Results: Tlfirty ascitic fluid infections (23 spontaneous bacterial peritonitis and 7 neutrocytic ascites) and 20
bacterascites were diagnosed. Conventional culture was positive in 10/30 ascitic fluid infections (33.3%), conventional blood culture bottles in 22/30 (73.3%) (p<0.01 compared to conventional culture) and BacT/ ALERT in 20/30 (66.6%) (p<0.05 compared to conventional culture, pNS compared to conventional blood culture bottles). The time elapsed for culture positivity was 43.4_+34.2 h for conventional blood culture bottles and 13.3_+9.2 h for BacT/ALERT (p<0.001). Thirteen of the 23 cases of spontaneous bacterial peritonitis (56.5%) were detected within the first 12 h with BacT/ALERT, as compared to only three (13%) with conventional blood culture bottles
PONTAr,rEOt;S bacterial peritonitis (SBP) is a severe and frequent complication in patients with cirrhosis and ascites (1). The incidence of SBP in patients with cirrhosis and ascites admitted to hospital varies from 7 to 27% (2-5), and the current mortality rate associated with an episode of SBP ranges between 20% and 45% (6). The diagnosis of SBP is based on a positive ascitic fluid culture and a polymorphonuclear (PMN) cell count ---250/mm 3, in the absence of an intraabdominal
source of infection (7,8). However, some patients with cirrhosis and ascites present a clinical and analytical picture of SBP without a positive ascitic fluid culture. Runyon & Hoefs defined this variant of SBP as culture-negative neutrocytic ascites (CNNA) (i.e. P M N count ->500/mm 3 with negative culture) (9). These authors suggested that culture negativity could be explained by the low sensitivity of the culture of ascitic fluid in agar plates and thioglycolate broth (conventional culture), due to the low concentration of bacteria in infected ascites, as occurs with blood in patients with bacteremia (9). Therefore, inoculation of ascitic fluid into blood culture bottles was introduced to improve culture sensitivity by Kammerer et al. (10). In 1987, in a study using a historical cohort, Runyon et al. (11) reported that direct inoculation of ascites into
S
Received 10 April," revised 14 October; accepted 21 October 1996
Correspondence: Dr. Germ~in Soriano, Liver Section, Department of Gastroenterology, Hospital de la Santa Creu i Sant Pau, S. Antoni M. Claret, 167, 08025 Barcelona, Spain. Tel: 34-3-2919139. Fax: 34-3-2919278.
(t,<0.03). Conclusion: The automated system BacT/ALERT provides an earlier microbiologic diagnosis of spontaneous bacterial peritonitis than conventional blood culture bottles with similar sensitivity. Key words: Ascitic fluid culture; BacT/ALERT; Cirrhosis; Infections; Spontaneous bacterial peritonitis..
839
•I. O r t i z e t aL
blood culture bottles at the bedside was a more sensitive method for detecting SBP (90%) than the conventional method (40%). Similar results have since been reported in several prospective studies (12-17). In spite of the improvement in culture sensitivity with this method, the time required for bacterial growth and its detection is usually more than 12-24 h, depending on the frequency of visual controls to check the macroscopic signs of bacterial growth (12). Considering the prognostic significance of early appropriate antibiotic treatment in patients with SBP (8,18), it would be interesting to find a method that could shorten the time for bacteria identification. BacT/ALERT is an automated microbial detection system based on the colorimetric detection of CO2 produced by growing microorganisms (19-21). It has proven to be faster than conventional blood culture bottles in patients with bacteremia (22-24). The aims of the present study were: (a) to prospectively evaluate the efficacy of BacT/ALERT system in diagnosing ascitic fluid infection in patients with cirrhosis, and the rapidity of bacterial growth detection using this system; and (b) to compare this method with conventional culture (CC) and conventional blood culture bottles (CBCB).
Materials and Methods All ascitic fluid samples from patients with cirrhosis hospitalized in our Department between September 1992 and May 1994 (n=1032) were included in the present study. A diagnostic paracentesis was performed in the following circumstances: (a) patients with cirrhosis and ascites at admission, (b) patients who developed signs or symptoms of SBP during hospitalization, i.e. fever, abdominal pain or encephalopathy, (c) to evaluate the response to treatment in patients with ascitic fluid infection, and d) patients submitted to a large volume paracentesis. Ascitic fluid samples were not included in the study when an inadequate amount of ascitic fluid was obtained. An ascitic fluid neutrophil count and a microbiological study were performed in all cases. An aliquot of ascitic fluid (3-5 ml) was sent for Gram's stain and CC, and 20 ml were inoculated at the bedside in blood culture bottles (10 ml in CBCB and 10 ml in BacT/ ALERT bottles). In the CC method 2-3 drops of uncentrifuged ascitic fluid were directly inoculated on chocolate, blood agar and Wilkins-Chalgren.plates, which were incubated at 35°C in aerobic (blood agar and chocolate) or anaerobic (Wilkins-Chalgren) atmosphere for 2 days and examined daily for visible growth. At the same time, 0.5-1 ml of uncentrifuged ascitic fluid was inoculated 840
in thioglycolate broth and also incubated at 35°C for 48 h. In the CBCB method, an aerobic and an anaerobic blood culture bottle (H6moline Performance DUO, Bio-M6rieux, France) were each inoculated with 5 ml of ascitic fluid at the bedside. Both bottles were incubated at 35°C for 7 days and examined for visible growth once a day. Each bottle was treated independently, and only bottles with visible growth were further processed. The time periods in which cultures became positive were recorded by the technicians. In the BacT/ALERT blood culture system (Organon Teknika Corp., Durham, NC), an aerobic and an anaerobic BacT/ALERT blood culture bottle were each inoculated with 5 ml of ascitic fluid at the bedside. The bottles were placed in the BacT/ALERT instrument and were processed according to the manufacturer's instructions. Each bottle was treated independently, and only bottles flagged as positive by the instrument were further processed. The time periods in which cultures became positive were automatically recorded. In the BacT/ALERT system the initial time was defined as" the moment when the bottles were placed in the instrument and, in the CBCB method as the moment when the bottles were placed in the incubator. The time elapsed for culture positivity in the BacT/ ALERT system was defined as the period between the initial time and the time when the system detected the bottle as positive. The time elapsed for culture positivity in the CBCB method was defined as the period between the initial time and the time when the bottle was detected as positive in the daily visual evaluation for macroscopic signs of growth. Isolated organisms were identified by standard methods (25). A positive ascitic fluid culture for Propionibacterium spp, corynebacteria, Bacillus spp or coagulase negative staphylococci was considered as a contamination. When patients presented fever (temperature ->38°C), 10 ml of blood (5 ml into each of the two bottles of BacT/ALERT) were also cultured at the time of paracentesis. We considered SBP to be present when the ascitic fluid neutrophil count was ->250/mm 3 with positive culture in the absence of any intraabdominal source of infection (8). CNNA was diagnosed when the ascitic fluid neutrophil count was ->500/mm 3 with a negative culture and without antibiotic therapy within 7 days or an alternative cause for elevated neutrophils in ascitic fluid (12). Bacterascites was defined as an ascitic fluid neutrophil count <250/mm 3 with a positive culture (26-28). Secondary bacterial peritonitis was diagnosed when the neutrophil count was ->250 cells/mm 3 with a
Early microbiologic
positive culture in the presence of an intraabdominal source of infection (8,29). The number of patients developing ascitic fluid infection or bacterascites while on selective intestinal decontamination with norfloxacin for prophylaxis of bacterial infection (30-32) was recorded. Patients with ascitic fluid infection were empirically treated with an intravenous third-generation cephalosporin (cefotaxime, 1-2 g/6 h) (6,18) and this treatment was changed when needed, according to the results of the cultures and antibiograms. Statistical analyses were performed by using the Student's t-test and the chi-square test with Yates' correction. Results are presented as mean_S.D. A p-value less than 0.05 was considered to be statistically significant.
Results Over a 21-month period, 34 ascitic fluid infections were diagnosed. Twenty-three ascitic fluid infections in 21 patients fulfilled SBP criteria and seven patients developed 7 CNNA episodes. Four patients were documented to have secondary bacterial peritonitis and were excluded from the analysis of the results. In the same period, 20 episodes of bacterascites were diagnosed in 12 patients. Moreover, in 23 additional cases with ascitic fluid neutrophil count <250/mm 3 the ascitic fluid culture grew bacteria that were considered as contaminant (14 coagulase negative staphylococci, 5 corynebacteria, 3 Propionibacterium acnes and 1 Bacillus spp) and were also excluded from the analysis of the results. In 10/23 (43.4%) episodes of SBP and 9/20 (45%) episodes of bacterascites, patients were on prophylactic selective intestinal decontamination with norfloxacin, since they belonged to groups at high risk of developing bacterial infection (30-32). Gram's stain was positive in only two ascitic fluid infections (6.6%), and in none of the bacterascites. The sensitivity of both CBCB (22/30, 73.3%) and BacT/ALERT (20/30, 66.6%) was significantly higher than CC (10/30, 33.3%) in diagnosing ascitic fluid infection (p<0.01 and p<0.05, respectively). No significant differences were observed between the two blood culture methods. In no cases did the CC grow bacteria when the blood culture bottle methods did not. The sensitivity of both CBCB (18/20, 90%) and BacT/ALERT (18/20, 90%) was significantly higher than CC (7/20, 35%) in diagnosing bacterascites @<0.005 in both cases). No significant differences were observed between the two blood culture systems. There were no statistical differences in sensitivity between aerobic and anaerobic bottles in any blood culture system (CBCB and BacT/ALERT), neither in as-
diagnosis of SBP
citic fluid infections (CBCB: aerobic 21/30, 70% vs anaerobic 20/30, 66.6%, p NS; BacT/ALERT: aerobic 19/ 30, 63.3% vs anaerobic 19/30, 63.3%, p NS) nor in bacterascites (CBCB: aerobic 18/20, 90% vs anaerobic 17/ 20, 85%, p NS; BacT/ALERT: aerobic 16/20, 80% vs anaerobic 17/20, 85%, p NS). There were no differences between the two blood culture methods in the detection of microorganisms considered as contaminant (12/23 in CBCB and 12/23 in BacT/ALERT, p NS). The isolated microorganisms are shown in Table 1. In this series gram-positive bacteria were predominant in patients with SBP and in those with bacterascites (69.2% and 61.9%, respectively). In three of the ascitic fluid infections (10%) and in one of the bacterascites (5%), two different bacteria were detected in the ascitic fluid culture. The bacteria responsible for SBP in patients submitted to selective intestinal decontamination with norfloxacin were: 6 Viridans streptococci, 2 Escherichia coli, 3 Enterococcus faecalis and 1 Staphylococcus aureus. The bacteria responsible for bacterascites in patients on prophylactic norfloxacin were: 7 Viridans streptococci, 1 Enterococcusfaecalis and 1 Citrobacter freundii. Blood cultures were performed in 12 patients with SBP and were positive in 6 (50%). In all cases, the microorganism isolated in blood was the same as that in ascitic fluid. Fig. 1 shows the time elapsed for culture positivity in the two blood culture bottle methods in patients with SBP and bacterascites. Interestingly, in patients with SBP the mean time elapsed was 43.4___34.2 (10-168) h
TABLE 1 Microorganismsisolatedbythethreeasciticfluidcul~remethods CC
CBCB
BA
SBP 7 E s c h e r i c h i a coli
3
7
7 Viridans group streptococci
3
6
7
6 2 2 1 1
Enterococcusfaecalis Streptococcus pneumoniae Streptococcus agalactiae Staphylococcus aureus Klebsiella pneumoniae
5 1 -
6 2 2 1 1
6 1 1 1
BACTERASCITES 11 Viridans group streptococci
6
11
11
1 1 -
4 1 1 1 1
4 1 1
4 1 1 1 1 1 1
E s c h e r i c h i a coli B a c t e r o i d e s spp Streptococcus pneumoniae Enterococcusfaecalis Campylobacter jejuni Citrobacterfreundii Klebsiella oxytoca
7
1 1
CC: conventional culture, CBCB: conventional blood culture bottles, BA: BacT/ALERT.
841
J. Ortiz et aL
150
100 -n O "1" e
50
N
• e•
CBCB
BA
CBCB BA I Bacterascites
I
Fig. 1. Time elapsed for positivity of ascitic fluid culture in spontaneous bacterial peritonitis (SBP) and bacterascites (p
for CBCB and 13.3___9.2 (6-48) h for BacT/ALERT (p<0.001) and in patients with bacterascites it was 30.5__.21.0 (11-76) h for CBCB and 18.6_ +_15.6 (8-78) h for BacT/ALERT (p=0.06). Moreover, the culture was positive within the first 12 h in 13 of the 23 cases of SBP (56.5%) with the BacT/ALERT system and in three cases with the CBCB system (13%) (p<0.03). In 6 of the 23 cases of SBP (26%) the initial empiric antibiotic therapy was changed according to the results of the culture and the in vitro susceptibility of isolated bacteria. All of the cases were caused by Enterococcus faecalis resistant to cefotaxime and were treated with amoxicillin/clavulanic acid, imipenem or vancomicin. Three of these patients had been submitted to previous prophylactic selective intestinal decontamination and two had previously been treated with multiple broad spectrum antibiotics. The cost of each set of blood culture bottles was USD 6.2 for CBCB and USD 10.4 for BacT/ALERT.
Discussion In agreement with previous data (10-17), this study confirms that bedside inoculation of ascitic fluid into 842
blood culture bottles is more effective than the CC method in the microbiologic diagnosis of SBP. On comparing two different methods of blood culture bottles (CBCB and BacT/ALERT), the automated system BacT/ALERT proved to be faster in detecting bacterial growth than CBCB, with similar sensitivity. All ascitic fluid cultures remained negative in 7 (23.3%) of the 30 episodes of suspected SBP. This incidence of CNNA is comparable with the results of previous studies using CBCB (11-15). Considering the diagnosis of SBP, both blood culture methods had a similar overall sensitivity (73.3% vs 66.6%, p NS). These results are comparable to those observed by Runyon (70%) with the inoculation of 10 ml (5 ml each bottle) in CBCB (12). The inoculation of larger amounts of ascitic fluid (10 ml per bottle) would probably result in higher sensitivity, as has been previously reported (12). Runyon et al. (12) found a higher sensitivity for the bottle containing thiol (anaerobic) (93%) than for the tryptic soy broth bottle (aerobic) (54%) in diagnosing SBP. Other authors have reported higher sensitivity for the tryptic soy broth blood culture bottle (13,15). In our study, there were no statistical differences in sensitivity between aerobic and anaerobic bottles for either of the two methods used. BacT/ALERT was faster in detecting bacterial growth than CBCB in patients with SBE More than 50% of the ascitic fluid cultures were positive in the first 12 h with BacT/ALERT and only 13% with CBCB. Therefore, BacT/ALERT provides an earlier microbiologic diagnosis of SBP that could determine an earlier appropriate antibiotic treatment. Yet in our study, the results of BacT/ALERT determined an earlier change in empiric antibiotic treatment in 26% of cases. In the present study, we have not evaluated the possible influence of BacT/ALERT on the morbidity and mortality of patients with SBP. However, BacT/ALERT could theoretically decrease the morbidity and mortality of this complication in patients not responding to initial empiric antibiotic treatment, since appropriate treatment seems to be a key factor in the outcome of patients with SBP (8,18). As in SBP, in bacterascites the two methods of blood culture bottles were more sensitive than CC (90% vs 35%, p<0.005 in both methods). There were no statistical differences in sensitivity between aerobic and anaerobic bottles in any method used. In bacterascites, BacT/ALERT was also faster than CBCB in detecting bacterial growth although without statistical significance. In patients with bacteremia, Alonso et al. (22) showed that BacT/ALERT has a non-significant higher
Early microbiologic diagnosis of SBP sensitivity for enterobacteria a n d lower sensitivity for gram-positive cocci c o m p a r e d to CBCB. In the current study we did n o t observe statistical differences in the type o f bacteria isolated between the two methods, alt h o u g h it should be n o t e d that m o s t bacteria t h a t did n o t g r o w with B a c T / A L E R T were gram-positive cocci. Gram-negative bacilli are the bacteria m o s t frequently isolated f r o m the ascitic fluid in patients with cirrhosis a n d SBP (6,8,11-15,18). Interestingly, we f o u n d that gram-positive bacteria were p r e d o m i n a n t in patients with SBP a n d in those with bacterascites (69.2% a n d 61.9%, respectively). This is p r o b a b l y related to the prophylactic treatment with oral norfloxacin, as the bacteria isolated in patients on prophylactic norfloxacin in the present study were mainly g r a m positive cocci. These results are in agreement with those previously observed in controlled trials (30-32), a n d would be explained by the antimicrobial activity o f norfloxacin, which is highly active against g r a m negative bacilli a n d less active against gram-positive cocci (30-34). In conclusion, the a u t o m a t e d system B a c T / A L E R T provides an earlier microbiologic diagnosis o f SBP t h a n C B C B with similar sensitivity. Therefore, BacT/ A L E R T allows an earlier a p p r o p r i a t e antibiotic treatm e n t in SBP that could improve m o r b i d i t y and m o r tality in patients with this complication n o t responding to the initial empiric antibiotic treatment.
Acknowledgements We are grateful to the medical staff, technical staff and nurses o f o u r D e p a r t m e n t s for their collaboration.
References i. Gilbert JA, Kamath PS. Spontaneous bacterial peritonitis: an update. Mayo Clin Proc 1995; 70: 365-70. 2. Pinzello G, Simonetti RG, Craxi A, Di Piazza S, Spano C, Pagliaro L. Spontaneous bacterial peritonitis: a prospective investigation in predominantly nonalcoholic cirrhotic patients. Hepatology 1983; 3: 545-9. 3. Caly WR, Strauss E. A prospective study of bacterial infections in patients with cirrhosis..t Hepatol 1993; 18: 353-8. 4. Runyon BA. Low-protein concentration ascitic fluid is predisposed to spontaneous bacterial peritonitis. Gastroenterology 1986; 91: 1343-6. 5. Almdal TP, Skinhoj P. Spontaneous bacterial peritonitis in cirrhosis. Incidence, diagnosis and prognosis. Scand J Gastroenterol 1987; 22: 295-300. 6. Arroyo V, Navasa M, Rimola A. Spontaneous bacterial peritonitis in liver cirrhosis: treatment and prophylaxis. Infection 1994; 22 (Suppl 3): 169-75. 7. Hoefs JC, Canawati HN, Sapico FL, Hopkins R, Weiner J, Montgomerie JZ. Spontaneous bacterial peritonitis. Hepatology 1982; 2: 399-407. 8. Hoefs JC, Runyon BA. Spontaneous bacterial peritonitis. DM 1985; 31: 1-48.
9. Runyon BA, Hoefs JC. Culture-negative neutrocytic ascites: a variant of spontaneous bacterial peritonitis. Hepatology 1984; 4: 1209-11. 10. Kammerer J, Dupeyron C, Vuillemin N, Leluan G, Fouet P. Apport des examens cytologiques et bact6riologiques du liquide d'ascite cirrhotique au diagnostic de p6ritonite bact6rienne. Med Chir Dig 1982; 11: 243-51. 11. Runyon BA, Umland ET, Merlin T. Inoculation of blood culture bottles with ascitic fluid. Arch Intern Med 1987; 147: 73-5. 12. Runyon BA, Canawati HN, Akriviadis EA. Optimization of ascitic fluid culture technique. Gastroenterology 1988; 95: 1351-5. 13. Bobadilla M, Sifuentes J, Garcia-Tsao G. Improved method for bacteriological diagnosis of spontaneous bacterial peritonitis. J Clin Microbiol 1989; 27: 2145-7. 14. Sainz S, Soriano G, Guarner C, Coil P, Vilardell E More on ascitic fluid culture technique. Hepatology 1989; 9:662-3 (letter). 15. Castellote J, Xiol X, Verdaguer R, Ribes J, Guardiola J, Gimenez A, Casais L. Comparison of two ascitic fluid culture methods in cirrhotic patients with spontaneous bacterial peritonitis. Am J Gastroenterol 1990; 85: 1605-8. 16. Siersema PD, De Marie S, Van Zeijl JH, Bac DJ, Wilson JHP. Blood culture bottles are superior to lysis-centrifugation tubes for bacteriological diagnosis of spontaneous bacterial peritonitis. J Clin Microbiol 1992; 30: 667-9. 17. Singh N, Rihs JD, Gayowsky T, Mieles L, Yu VL. Improved detection of spontaneous bacterial peritonitis with Bactec as compared with conventional culture methods. A prospective study. Diagn Microbiol Infect Dis 1994; 1: 1-4. 18. Guarner C, Runyon BA. Spontaneous bacterial peritonitis: pathogenesis, diagnosis and management. Gastroenterologist 1995; 3: 311-28. 19. Thorpe TC, Wilson ML, Turner JE, DiGuiseppi JL, Willert M, Mirrett S, Relier SB. BacT/ALERT: an automated colorimetric microbial detection system. J Clin Microbiol 1990; 28: 1608-12. 20. Wilson ML, Weinstein MP, Reimer LG, Mirrett S, Relier LB. Controlled comparison of the BacT/ALERT and BACTEC 600/730 nonradiometric blood culture systems. J Clin Microbiol 1992; 30: 323-9. 21. Smith JA, Bryce EA, Ngui-Yen JH, Roberts FJ. Comparison of BACTEC 9240 and BacT/ALERT blood culture systems in an adult hospital. J Clin Microbiol 1995; 33: 1905-8. 22. Alonso C, Rello J, Mirelis B, Pericas R, Navarro E Prats G. Comparaci6n del sistema automatizado de incubaci6n y lectura de hemocultivos BacT/ALERT con un sistema convencional. Enferm Infecc Microbiol Clin 1995; 13: 17-22. 23. Rohner P, Pepey B, Auckenthaler R. Comparison of BacT/ ALERT with signal blood culture system. J Clin Microbiol 1995; 33: 313-7. 24. Hellinger WC, Cawley J J, Alvarez S, Hogan SE Harmsen WS, Ilstrup DM, Cockerill III FR. Clinical comparison of the isolator and BacT/ALERT aerobic blood culture systems. J Clin Microbiol 1995; 33: 1787-90. 25. Murray PR. Manual of clinical microbiology, 6th edn. Washington,D.C.: American Society for Microbiology Press, 1995. 26. Runyon BA. Monomicrobial nonneutrocytic bacterascites: a variant of spontaneous bacterial peritonitis. Hepatology 1990; 12: 710-5. 27. Pelletier G, Lesur G, Ink O, Hagege H, Attali P, Buffet C, Etienne JP. Asymptomatic bacterascites: is it spontaneous bacterial peritonitis? Hepatology 1991; 14: 112-5.
843
J. Ortiz et al.
28. Chu CM, Chang KE Liaw YE Prevalence and prognostic significance of bacterascites in cirrhosis with ascites. Dig Dis Sci 1995; 40: 561-5. 29. Runyon BA, Hoefs JC. Ascitic fuid analysis in the differentiation of spontaneous bacterial peritonitis from gastrointestinal tract perforation into ascitic fluid. Hepatotogy 1984; 4: 447-50. 30. Gin6s P, Rimola A, Planas R, Vargas V, Marco E Almela M, Forn6 M, Miranda ML, Llach J, Salmer6n JM, Esteve M, Marqu6s M, Jimenez de Anta MT, Arroyo V, Rod6s J. Norfloxacin prevents spontaneous bacterial peritonitis recurrence in cirrhosis: results of a double-blind, placebo controlled trial. Hepatology 1990; 12: 716-24. 31. Soriano G, Guarner C, Teixid6 M, Such J, Barrios J, En-
844
riquez J, Vilardell J. Selective intestinal decontamination prevents spontaneous bacterial peritonitis. Gastroenterology 1991; 100: 477-81. 32. Soriano G, Guarner C, Tom,is A, Villanueva C, Torras X, Gonz~lez D, Sainz S, Anguera A, Cuss6 X, Balanz6 J, Vilardell E Norfloxacin prevents bacterial infection in cirrhotics with gastrointestinal hemorrhage. Gastroenterology 1992; 103: 1267-72. 33. Tom,is A, Soriano G, Guarner C, Portorreal R, Novella MT, Vilardell E Hospital-acquired bacterial infections in patients with cirrhosis undergoing selective intestinal decontamination (letter). J Hepatot 1993; 18: 262-3. 34. Nord CE. Effect of new quinolones on the human gastrointestinal microflora. Rex,Infect Dis 1988; 10 (suppl l): 5193-6.
Journal of Hepatology 1997; 26:839-844 Printed in Denmark • All rights reserved Munksgaard. Copenhagen
Journal of Hepatology ISSN 0168-8278
Early microbiologic diagnosis of spontaneous bacterial peritonitis with BacT/ALEgF Jordi Ortiz, Germfin Soriano, Pere Coil 1, Maria Teresa Novella, Roser Pericas 1, Miriam S/tbat, Ferran Sfinchez 1, Carlos Guarner, Guillem Prats I and Francisco Vilardell Departments o f Gastroenterology and IMicrobiology, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain
Background~Aims: Inoculation of ascitic fluid into conventional blood culture bottles is more sensitive than conventional culture in the diagnosis of spontaneous bacterial peritonitis. BacT/ALERT is an automated colorimetric microbial detection system that has been shown to be faster than conventional blood culture bottles in the diagnosis of bacteremia. The aim of the study was to compare the BacT/ ALERT system with the conventional culture and the conventional blood culture bottles method in the diagnosis of spontaneous bacterial peritonitis. Methods: All the ascitic fluid samples from patients with cirrhosis hospitalized in our Department between September 1992 and May 1994 (n=1032) were prospectively evaluated. In all cases, an aliquot of ascitic fluid was sent for Gram's stain and conventional culture, and 20 ml were inoculated at the bedside into blood culture bottles: 10 ml into conventional blood culture bottles and 10 mi into BacT/ALERT. Results: Tlfirty ascitic fluid infections (23 spontaneous bacterial peritonitis and 7 neutrocytic ascites) and 20
bacterascites were diagnosed. Conventional culture was positive in 10/30 ascitic fluid infections (33.3%), conventional blood culture bottles in 22/30 (73.3%) (p<0.01 compared to conventional culture) and BacT/ ALERT in 20/30 (66.6%) (p<0.05 compared to conventional culture, pNS compared to conventional blood culture bottles). The time elapsed for culture positivity was 43.4_+34.2 h for conventional blood culture bottles and 13.3_+9.2 h for BacT/ALERT (p<0.001). Thirteen of the 23 cases of spontaneous bacterial peritonitis (56.5%) were detected within the first 12 h with BacT/ALERT, as compared to only three (13%) with conventional blood culture bottles
PONTAr,rEOt;S bacterial peritonitis (SBP) is a severe and frequent complication in patients with cirrhosis and ascites (1). The incidence of SBP in patients with cirrhosis and ascites admitted to hospital varies from 7 to 27% (2-5), and the current mortality rate associated with an episode of SBP ranges between 20% and 45% (6). The diagnosis of SBP is based on a positive ascitic fluid culture and a polymorphonuclear (PMN) cell count ---250/mm 3, in the absence of an intraabdominal
source of infection (7,8). However, some patients with cirrhosis and ascites present a clinical and analytical picture of SBP without a positive ascitic fluid culture. Runyon & Hoefs defined this variant of SBP as culture-negative neutrocytic ascites (CNNA) (i.e. P M N count ->500/mm 3 with negative culture) (9). These authors suggested that culture negativity could be explained by the low sensitivity of the culture of ascitic fluid in agar plates and thioglycolate broth (conventional culture), due to the low concentration of bacteria in infected ascites, as occurs with blood in patients with bacteremia (9). Therefore, inoculation of ascitic fluid into blood culture bottles was introduced to improve culture sensitivity by Kammerer et al. (10). In 1987, in a study using a historical cohort, Runyon et al. (11) reported that direct inoculation of ascites into
S
Received 10 April," revised 14 October; accepted 21 October 1996
Correspondence: Dr. Germ~in Soriano, Liver Section, Department of Gastroenterology, Hospital de la Santa Creu i Sant Pau, S. Antoni M. Claret, 167, 08025 Barcelona, Spain. Tel: 34-3-2919139. Fax: 34-3-2919278.
(t,<0.03). Conclusion: The automated system BacT/ALERT provides an earlier microbiologic diagnosis of spontaneous bacterial peritonitis than conventional blood culture bottles with similar sensitivity. Key words: Ascitic fluid culture; BacT/ALERT; Cirrhosis; Infections; Spontaneous bacterial peritonitis..
839
•I. O r t i z e t aL
blood culture bottles at the bedside was a more sensitive method for detecting SBP (90%) than the conventional method (40%). Similar results have since been reported in several prospective studies (12-17). In spite of the improvement in culture sensitivity with this method, the time required for bacterial growth and its detection is usually more than 12-24 h, depending on the frequency of visual controls to check the macroscopic signs of bacterial growth (12). Considering the prognostic significance of early appropriate antibiotic treatment in patients with SBP (8,18), it would be interesting to find a method that could shorten the time for bacteria identification. BacT/ALERT is an automated microbial detection system based on the colorimetric detection of CO2 produced by growing microorganisms (19-21). It has proven to be faster than conventional blood culture bottles in patients with bacteremia (22-24). The aims of the present study were: (a) to prospectively evaluate the efficacy of BacT/ALERT system in diagnosing ascitic fluid infection in patients with cirrhosis, and the rapidity of bacterial growth detection using this system; and (b) to compare this method with conventional culture (CC) and conventional blood culture bottles (CBCB).
Materials and Methods All ascitic fluid samples from patients with cirrhosis hospitalized in our Department between September 1992 and May 1994 (n=1032) were included in the present study. A diagnostic paracentesis was performed in the following circumstances: (a) patients with cirrhosis and ascites at admission, (b) patients who developed signs or symptoms of SBP during hospitalization, i.e. fever, abdominal pain or encephalopathy, (c) to evaluate the response to treatment in patients with ascitic fluid infection, and d) patients submitted to a large volume paracentesis. Ascitic fluid samples were not included in the study when an inadequate amount of ascitic fluid was obtained. An ascitic fluid neutrophil count and a microbiological study were performed in all cases. An aliquot of ascitic fluid (3-5 ml) was sent for Gram's stain and CC, and 20 ml were inoculated at the bedside in blood culture bottles (10 ml in CBCB and 10 ml in BacT/ ALERT bottles). In the CC method 2-3 drops of uncentrifuged ascitic fluid were directly inoculated on chocolate, blood agar and Wilkins-Chalgren.plates, which were incubated at 35°C in aerobic (blood agar and chocolate) or anaerobic (Wilkins-Chalgren) atmosphere for 2 days and examined daily for visible growth. At the same time, 0.5-1 ml of uncentrifuged ascitic fluid was inoculated 840
in thioglycolate broth and also incubated at 35°C for 48 h. In the CBCB method, an aerobic and an anaerobic blood culture bottle (H6moline Performance DUO, Bio-M6rieux, France) were each inoculated with 5 ml of ascitic fluid at the bedside. Both bottles were incubated at 35°C for 7 days and examined for visible growth once a day. Each bottle was treated independently, and only bottles with visible growth were further processed. The time periods in which cultures became positive were recorded by the technicians. In the BacT/ALERT blood culture system (Organon Teknika Corp., Durham, NC), an aerobic and an anaerobic BacT/ALERT blood culture bottle were each inoculated with 5 ml of ascitic fluid at the bedside. The bottles were placed in the BacT/ALERT instrument and were processed according to the manufacturer's instructions. Each bottle was treated independently, and only bottles flagged as positive by the instrument were further processed. The time periods in which cultures became positive were automatically recorded. In the BacT/ALERT system the initial time was defined as" the moment when the bottles were placed in the instrument and, in the CBCB method as the moment when the bottles were placed in the incubator. The time elapsed for culture positivity in the BacT/ ALERT system was defined as the period between the initial time and the time when the system detected the bottle as positive. The time elapsed for culture positivity in the CBCB method was defined as the period between the initial time and the time when the bottle was detected as positive in the daily visual evaluation for macroscopic signs of growth. Isolated organisms were identified by standard methods (25). A positive ascitic fluid culture for Propionibacterium spp, corynebacteria, Bacillus spp or coagulase negative staphylococci was considered as a contamination. When patients presented fever (temperature ->38°C), 10 ml of blood (5 ml into each of the two bottles of BacT/ALERT) were also cultured at the time of paracentesis. We considered SBP to be present when the ascitic fluid neutrophil count was ->250/mm 3 with positive culture in the absence of any intraabdominal source of infection (8). CNNA was diagnosed when the ascitic fluid neutrophil count was ->500/mm 3 with a negative culture and without antibiotic therapy within 7 days or an alternative cause for elevated neutrophils in ascitic fluid (12). Bacterascites was defined as an ascitic fluid neutrophil count <250/mm 3 with a positive culture (26-28). Secondary bacterial peritonitis was diagnosed when the neutrophil count was ->250 cells/mm 3 with a
Early microbiologic
positive culture in the presence of an intraabdominal source of infection (8,29). The number of patients developing ascitic fluid infection or bacterascites while on selective intestinal decontamination with norfloxacin for prophylaxis of bacterial infection (30-32) was recorded. Patients with ascitic fluid infection were empirically treated with an intravenous third-generation cephalosporin (cefotaxime, 1-2 g/6 h) (6,18) and this treatment was changed when needed, according to the results of the cultures and antibiograms. Statistical analyses were performed by using the Student's t-test and the chi-square test with Yates' correction. Results are presented as mean_S.D. A p-value less than 0.05 was considered to be statistically significant.
Results Over a 21-month period, 34 ascitic fluid infections were diagnosed. Twenty-three ascitic fluid infections in 21 patients fulfilled SBP criteria and seven patients developed 7 CNNA episodes. Four patients were documented to have secondary bacterial peritonitis and were excluded from the analysis of the results. In the same period, 20 episodes of bacterascites were diagnosed in 12 patients. Moreover, in 23 additional cases with ascitic fluid neutrophil count <250/mm 3 the ascitic fluid culture grew bacteria that were considered as contaminant (14 coagulase negative staphylococci, 5 corynebacteria, 3 Propionibacterium acnes and 1 Bacillus spp) and were also excluded from the analysis of the results. In 10/23 (43.4%) episodes of SBP and 9/20 (45%) episodes of bacterascites, patients were on prophylactic selective intestinal decontamination with norfloxacin, since they belonged to groups at high risk of developing bacterial infection (30-32). Gram's stain was positive in only two ascitic fluid infections (6.6%), and in none of the bacterascites. The sensitivity of both CBCB (22/30, 73.3%) and BacT/ALERT (20/30, 66.6%) was significantly higher than CC (10/30, 33.3%) in diagnosing ascitic fluid infection (p<0.01 and p<0.05, respectively). No significant differences were observed between the two blood culture methods. In no cases did the CC grow bacteria when the blood culture bottle methods did not. The sensitivity of both CBCB (18/20, 90%) and BacT/ALERT (18/20, 90%) was significantly higher than CC (7/20, 35%) in diagnosing bacterascites @<0.005 in both cases). No significant differences were observed between the two blood culture systems. There were no statistical differences in sensitivity between aerobic and anaerobic bottles in any blood culture system (CBCB and BacT/ALERT), neither in as-
diagnosis of SBP
citic fluid infections (CBCB: aerobic 21/30, 70% vs anaerobic 20/30, 66.6%, p NS; BacT/ALERT: aerobic 19/ 30, 63.3% vs anaerobic 19/30, 63.3%, p NS) nor in bacterascites (CBCB: aerobic 18/20, 90% vs anaerobic 17/ 20, 85%, p NS; BacT/ALERT: aerobic 16/20, 80% vs anaerobic 17/20, 85%, p NS). There were no differences between the two blood culture methods in the detection of microorganisms considered as contaminant (12/23 in CBCB and 12/23 in BacT/ALERT, p NS). The isolated microorganisms are shown in Table 1. In this series gram-positive bacteria were predominant in patients with SBP and in those with bacterascites (69.2% and 61.9%, respectively). In three of the ascitic fluid infections (10%) and in one of the bacterascites (5%), two different bacteria were detected in the ascitic fluid culture. The bacteria responsible for SBP in patients submitted to selective intestinal decontamination with norfloxacin were: 6 Viridans streptococci, 2 Escherichia coli, 3 Enterococcus faecalis and 1 Staphylococcus aureus. The bacteria responsible for bacterascites in patients on prophylactic norfloxacin were: 7 Viridans streptococci, 1 Enterococcusfaecalis and 1 Citrobacter freundii. Blood cultures were performed in 12 patients with SBP and were positive in 6 (50%). In all cases, the microorganism isolated in blood was the same as that in ascitic fluid. Fig. 1 shows the time elapsed for culture positivity in the two blood culture bottle methods in patients with SBP and bacterascites. Interestingly, in patients with SBP the mean time elapsed was 43.4___34.2 (10-168) h
TABLE 1 Microorganismsisolatedbythethreeasciticfluidcul~remethods CC
CBCB
BA
SBP 7 E s c h e r i c h i a coli
3
7
7 Viridans group streptococci
3
6
7
6 2 2 1 1
Enterococcusfaecalis Streptococcus pneumoniae Streptococcus agalactiae Staphylococcus aureus Klebsiella pneumoniae
5 1 -
6 2 2 1 1
6 1 1 1
BACTERASCITES 11 Viridans group streptococci
6
11
11
1 1 -
4 1 1 1 1
4 1 1
4 1 1 1 1 1 1
E s c h e r i c h i a coli B a c t e r o i d e s spp Streptococcus pneumoniae Enterococcusfaecalis Campylobacter jejuni Citrobacterfreundii Klebsiella oxytoca
7
1 1
CC: conventional culture, CBCB: conventional blood culture bottles, BA: BacT/ALERT.
841
J. Ortiz et aL
150
100 -n O "1" e
50
N
• e•
CBCB
BA
CBCB BA I Bacterascites
I
Fig. 1. Time elapsed for positivity of ascitic fluid culture in spontaneous bacterial peritonitis (SBP) and bacterascites (p
for CBCB and 13.3___9.2 (6-48) h for BacT/ALERT (p<0.001) and in patients with bacterascites it was 30.5__.21.0 (11-76) h for CBCB and 18.6_ +_15.6 (8-78) h for BacT/ALERT (p=0.06). Moreover, the culture was positive within the first 12 h in 13 of the 23 cases of SBP (56.5%) with the BacT/ALERT system and in three cases with the CBCB system (13%) (p<0.03). In 6 of the 23 cases of SBP (26%) the initial empiric antibiotic therapy was changed according to the results of the culture and the in vitro susceptibility of isolated bacteria. All of the cases were caused by Enterococcus faecalis resistant to cefotaxime and were treated with amoxicillin/clavulanic acid, imipenem or vancomicin. Three of these patients had been submitted to previous prophylactic selective intestinal decontamination and two had previously been treated with multiple broad spectrum antibiotics. The cost of each set of blood culture bottles was USD 6.2 for CBCB and USD 10.4 for BacT/ALERT.
Discussion In agreement with previous data (10-17), this study confirms that bedside inoculation of ascitic fluid into 842
blood culture bottles is more effective than the CC method in the microbiologic diagnosis of SBP. On comparing two different methods of blood culture bottles (CBCB and BacT/ALERT), the automated system BacT/ALERT proved to be faster in detecting bacterial growth than CBCB, with similar sensitivity. All ascitic fluid cultures remained negative in 7 (23.3%) of the 30 episodes of suspected SBP. This incidence of CNNA is comparable with the results of previous studies using CBCB (11-15). Considering the diagnosis of SBP, both blood culture methods had a similar overall sensitivity (73.3% vs 66.6%, p NS). These results are comparable to those observed by Runyon (70%) with the inoculation of 10 ml (5 ml each bottle) in CBCB (12). The inoculation of larger amounts of ascitic fluid (10 ml per bottle) would probably result in higher sensitivity, as has been previously reported (12). Runyon et al. (12) found a higher sensitivity for the bottle containing thiol (anaerobic) (93%) than for the tryptic soy broth bottle (aerobic) (54%) in diagnosing SBP. Other authors have reported higher sensitivity for the tryptic soy broth blood culture bottle (13,15). In our study, there were no statistical differences in sensitivity between aerobic and anaerobic bottles for either of the two methods used. BacT/ALERT was faster in detecting bacterial growth than CBCB in patients with SBE More than 50% of the ascitic fluid cultures were positive in the first 12 h with BacT/ALERT and only 13% with CBCB. Therefore, BacT/ALERT provides an earlier microbiologic diagnosis of SBP that could determine an earlier appropriate antibiotic treatment. Yet in our study, the results of BacT/ALERT determined an earlier change in empiric antibiotic treatment in 26% of cases. In the present study, we have not evaluated the possible influence of BacT/ALERT on the morbidity and mortality of patients with SBP. However, BacT/ALERT could theoretically decrease the morbidity and mortality of this complication in patients not responding to initial empiric antibiotic treatment, since appropriate treatment seems to be a key factor in the outcome of patients with SBP (8,18). As in SBP, in bacterascites the two methods of blood culture bottles were more sensitive than CC (90% vs 35%, p<0.005 in both methods). There were no statistical differences in sensitivity between aerobic and anaerobic bottles in any method used. In bacterascites, BacT/ALERT was also faster than CBCB in detecting bacterial growth although without statistical significance. In patients with bacteremia, Alonso et al. (22) showed that BacT/ALERT has a non-significant higher
Early microbiologic diagnosis of SBP sensitivity for enterobacteria a n d lower sensitivity for gram-positive cocci c o m p a r e d to CBCB. In the current study we did n o t observe statistical differences in the type o f bacteria isolated between the two methods, alt h o u g h it should be n o t e d that m o s t bacteria t h a t did n o t g r o w with B a c T / A L E R T were gram-positive cocci. Gram-negative bacilli are the bacteria m o s t frequently isolated f r o m the ascitic fluid in patients with cirrhosis a n d SBP (6,8,11-15,18). Interestingly, we f o u n d that gram-positive bacteria were p r e d o m i n a n t in patients with SBP a n d in those with bacterascites (69.2% a n d 61.9%, respectively). This is p r o b a b l y related to the prophylactic treatment with oral norfloxacin, as the bacteria isolated in patients on prophylactic norfloxacin in the present study were mainly g r a m positive cocci. These results are in agreement with those previously observed in controlled trials (30-32), a n d would be explained by the antimicrobial activity o f norfloxacin, which is highly active against g r a m negative bacilli a n d less active against gram-positive cocci (30-34). In conclusion, the a u t o m a t e d system B a c T / A L E R T provides an earlier microbiologic diagnosis o f SBP t h a n C B C B with similar sensitivity. Therefore, BacT/ A L E R T allows an earlier a p p r o p r i a t e antibiotic treatm e n t in SBP that could improve m o r b i d i t y and m o r tality in patients with this complication n o t responding to the initial empiric antibiotic treatment.
Acknowledgements We are grateful to the medical staff, technical staff and nurses o f o u r D e p a r t m e n t s for their collaboration.
References i. Gilbert JA, Kamath PS. Spontaneous bacterial peritonitis: an update. Mayo Clin Proc 1995; 70: 365-70. 2. Pinzello G, Simonetti RG, Craxi A, Di Piazza S, Spano C, Pagliaro L. Spontaneous bacterial peritonitis: a prospective investigation in predominantly nonalcoholic cirrhotic patients. Hepatology 1983; 3: 545-9. 3. Caly WR, Strauss E. A prospective study of bacterial infections in patients with cirrhosis..t Hepatol 1993; 18: 353-8. 4. Runyon BA. Low-protein concentration ascitic fluid is predisposed to spontaneous bacterial peritonitis. Gastroenterology 1986; 91: 1343-6. 5. Almdal TP, Skinhoj P. Spontaneous bacterial peritonitis in cirrhosis. Incidence, diagnosis and prognosis. Scand J Gastroenterol 1987; 22: 295-300. 6. Arroyo V, Navasa M, Rimola A. Spontaneous bacterial peritonitis in liver cirrhosis: treatment and prophylaxis. Infection 1994; 22 (Suppl 3): 169-75. 7. Hoefs JC, Canawati HN, Sapico FL, Hopkins R, Weiner J, Montgomerie JZ. Spontaneous bacterial peritonitis. Hepatology 1982; 2: 399-407. 8. Hoefs JC, Runyon BA. Spontaneous bacterial peritonitis. DM 1985; 31: 1-48.
9. Runyon BA, Hoefs JC. Culture-negative neutrocytic ascites: a variant of spontaneous bacterial peritonitis. Hepatology 1984; 4: 1209-11. 10. Kammerer J, Dupeyron C, Vuillemin N, Leluan G, Fouet P. Apport des examens cytologiques et bact6riologiques du liquide d'ascite cirrhotique au diagnostic de p6ritonite bact6rienne. Med Chir Dig 1982; 11: 243-51. 11. Runyon BA, Umland ET, Merlin T. Inoculation of blood culture bottles with ascitic fluid. Arch Intern Med 1987; 147: 73-5. 12. Runyon BA, Canawati HN, Akriviadis EA. Optimization of ascitic fluid culture technique. Gastroenterology 1988; 95: 1351-5. 13. Bobadilla M, Sifuentes J, Garcia-Tsao G. Improved method for bacteriological diagnosis of spontaneous bacterial peritonitis. J Clin Microbiol 1989; 27: 2145-7. 14. Sainz S, Soriano G, Guarner C, Coil P, Vilardell E More on ascitic fluid culture technique. Hepatology 1989; 9:662-3 (letter). 15. Castellote J, Xiol X, Verdaguer R, Ribes J, Guardiola J, Gimenez A, Casais L. Comparison of two ascitic fluid culture methods in cirrhotic patients with spontaneous bacterial peritonitis. Am J Gastroenterol 1990; 85: 1605-8. 16. Siersema PD, De Marie S, Van Zeijl JH, Bac DJ, Wilson JHP. Blood culture bottles are superior to lysis-centrifugation tubes for bacteriological diagnosis of spontaneous bacterial peritonitis. J Clin Microbiol 1992; 30: 667-9. 17. Singh N, Rihs JD, Gayowsky T, Mieles L, Yu VL. Improved detection of spontaneous bacterial peritonitis with Bactec as compared with conventional culture methods. A prospective study. Diagn Microbiol Infect Dis 1994; 1: 1-4. 18. Guarner C, Runyon BA. Spontaneous bacterial peritonitis: pathogenesis, diagnosis and management. Gastroenterologist 1995; 3: 311-28. 19. Thorpe TC, Wilson ML, Turner JE, DiGuiseppi JL, Willert M, Mirrett S, Relier SB. BacT/ALERT: an automated colorimetric microbial detection system. J Clin Microbiol 1990; 28: 1608-12. 20. Wilson ML, Weinstein MP, Reimer LG, Mirrett S, Relier LB. Controlled comparison of the BacT/ALERT and BACTEC 600/730 nonradiometric blood culture systems. J Clin Microbiol 1992; 30: 323-9. 21. Smith JA, Bryce EA, Ngui-Yen JH, Roberts FJ. Comparison of BACTEC 9240 and BacT/ALERT blood culture systems in an adult hospital. J Clin Microbiol 1995; 33: 1905-8. 22. Alonso C, Rello J, Mirelis B, Pericas R, Navarro E Prats G. Comparaci6n del sistema automatizado de incubaci6n y lectura de hemocultivos BacT/ALERT con un sistema convencional. Enferm Infecc Microbiol Clin 1995; 13: 17-22. 23. Rohner P, Pepey B, Auckenthaler R. Comparison of BacT/ ALERT with signal blood culture system. J Clin Microbiol 1995; 33: 313-7. 24. Hellinger WC, Cawley J J, Alvarez S, Hogan SE Harmsen WS, Ilstrup DM, Cockerill III FR. Clinical comparison of the isolator and BacT/ALERT aerobic blood culture systems. J Clin Microbiol 1995; 33: 1787-90. 25. Murray PR. Manual of clinical microbiology, 6th edn. Washington,D.C.: American Society for Microbiology Press, 1995. 26. Runyon BA. Monomicrobial nonneutrocytic bacterascites: a variant of spontaneous bacterial peritonitis. Hepatology 1990; 12: 710-5. 27. Pelletier G, Lesur G, Ink O, Hagege H, Attali P, Buffet C, Etienne JP. Asymptomatic bacterascites: is it spontaneous bacterial peritonitis? Hepatology 1991; 14: 112-5.
843
J. Ortiz et al.
28. Chu CM, Chang KE Liaw YE Prevalence and prognostic significance of bacterascites in cirrhosis with ascites. Dig Dis Sci 1995; 40: 561-5. 29. Runyon BA, Hoefs JC. Ascitic fuid analysis in the differentiation of spontaneous bacterial peritonitis from gastrointestinal tract perforation into ascitic fluid. Hepatotogy 1984; 4: 447-50. 30. Gin6s P, Rimola A, Planas R, Vargas V, Marco E Almela M, Forn6 M, Miranda ML, Llach J, Salmer6n JM, Esteve M, Marqu6s M, Jimenez de Anta MT, Arroyo V, Rod6s J. Norfloxacin prevents spontaneous bacterial peritonitis recurrence in cirrhosis: results of a double-blind, placebo controlled trial. Hepatology 1990; 12: 716-24. 31. Soriano G, Guarner C, Teixid6 M, Such J, Barrios J, En-
844
riquez J, Vilardell J. Selective intestinal decontamination prevents spontaneous bacterial peritonitis. Gastroenterology 1991; 100: 477-81. 32. Soriano G, Guarner C, Tom,is A, Villanueva C, Torras X, Gonz~lez D, Sainz S, Anguera A, Cuss6 X, Balanz6 J, Vilardell E Norfloxacin prevents bacterial infection in cirrhotics with gastrointestinal hemorrhage. Gastroenterology 1992; 103: 1267-72. 33. Tom,is A, Soriano G, Guarner C, Portorreal R, Novella MT, Vilardell E Hospital-acquired bacterial infections in patients with cirrhosis undergoing selective intestinal decontamination (letter). J Hepatot 1993; 18: 262-3. 34. Nord CE. Effect of new quinolones on the human gastrointestinal microflora. Rex,Infect Dis 1988; 10 (suppl l): 5193-6.