Norfloxacin Prevents Spontaneous Bacterial Peritonitis Recurrence in Cirrhosis: Results of a Double-blind, Placebo-controlled Trial PERE GINl%,' ANTON1 RIMOLA,' RAMON PLANAS,4VICTOR VARGAS; FRANCESC MARCO,' MANUEL ALMELA,' MONTSERRAT FORNI~,~ MARIA LUISAMIRANDA,7 JOSEP LLACH,'JOAN MANUEL SALMERON,' MARIA ESTEVE,~ JOSEP MARIA MARQUl&,3 MARIA TERESA JIMI~NEZ DE ANTA,* VICENTE ARROYO' AND JOAN RODI~S'

'Liver Unit and 'Microbiological and 3Toxicology Laboratories, Hospital Clinic i Provincial of Barcelona; *Gastroenterology Unit; Hospital Germans Trim i Pujol of Badalona; 5Liver Unit, Hospital de la Val1 d'Hebrc5 of Barcelona. 'Department of Internal Medicine; Hospital Mutua of Terrassa and 7Department of Internal Medicine; Residencia Virgen del Rocio of Sevilla, Spain

Eighty cirrhotic patients who had recovered from an episode of spontaneous bacterial peritonitis were included in a multicenter, double-blind trial aimed at comparing long-term norfloxacin administration (400 mglday; 40 patients) vs. placebo (40 patients) in the prevention of spontaneous bacterial peritonitis recurrence. At entry, both groups were similar with respect to clinical and laboratorydata, ascitic fluid protein and polymorphonuclear concentrations, number of previous episodesof spontaneousbacterial peritonitis and causative organisms of the index spontaneous bacterial peritonitis. Norfloxacin administration produced a selective intestinal decontamination (elimination of aerobic gram-negative bacilli from the fecal flora without significant changes in other microorganisms) throughout the study in six patients in whom the effect of norfloxacin on the fecal flora was periodically assessed. Fourteen patients from the placebo group (35%) and five from the norfloxacin group (12%) developed spontaneous bacterial peritonitis recurrence during follow-up ($ = 5.97; p = 0.014) (mean follow-up period = 6.4 f 0.6 mo; range = 1 to 19 mo). Ten of the 14 spontaneous bacterial peritonitis recurrences in the placebo group and only one of the five spontaneous bacterial peritonitis recurrences in the norfloxacin group were caused by aerobic gramnegative bacilli ($ = 8.87; p = 0.0029). The overall probability of spontaneousbacterial peritonitis recurrence at 1 yr of follow-up was 20% in the norfloxacin group and 68% in the placebo group (p = 0.0063) and the probability of spontaneous bacterial peritonitis recurrence caused by aerobic gram-negative bacilli at 1 yr of follow-up was 3% and 60%, respectively (p = 0.0013). Only one patient treated with norfloxacin experienced side effects related to treatment (oral and

esophageal candidiasis). These results indicate that long-term selective intestinal decontamination with norfloxacin is an effective and safe measure to prevent spontaneous bacterial peritonitis recurrence caused by aerobic gram-negative bacilli in cirrhosis. (HEPATOLUCY 1990;12:716-724.)

Spontaneous bacterial peritonitis (SBP), a frequent and severe complication of cirrhosis, is present at admission or develops during hospitalization in 8% to 25% of cirrhotic patients with ascites admitted to a general hospital (1-5).The in-hospital mortality rate of cirrhotic patients with SBP has been reported to range between 30% and 50% (6-81, although in many cases death occurs after the resolution of SBP. Because most episodes of SBP are caused by organisms usually present in the intestinal flora, particularly aerobic (facultative) gram-negative bacilli (9), the most accepted pathogenic mechanism of SBP is the passage of intestinal bacteria into the general circulation and then into the ascitic fluid (9,101. An outstanding feature of cirrhotic patients who have recovered from SBP is that they are predisposed to experience new episodes of this complication during the course of the disease. A recent study has estimated that the l-yr probability rate of SBP recurrence in these patients is 69% (11). Selective intestinal decontamination (SID), elimination of aerobic gram-negative bacilli from the intestinal flora, but preserving the remaining aerobic and the anaerobic bacteria (12, 13) with oral nonabsorbable or poorly absorbable antibiotics, is a well-recognized effective measure in preventing infections caused by gram-negative bacilli in patients with granulocytopenia (14-16). The administration of oral nonabsorbable antibiotics also reduces the incidence of infections by Received November 27, 1989; accepted March 7, 1990. This work was supported by a grant from the Fondo de Investigaciones gram-negative organisms in cirrhotic patients admitted Sanitariaa de la Seguridad Social (FISS 86/465) and by the Fundacio Catalana to the hospital with gastrointestinal hemorrhage (17). per a 1'Estudi de les Malalties del Fetge. Therefore SID could be useful in preventing SBP Address reprint requests to: Dr. Pere Gin&, Liver Unit, Hospital Clinic i recurrence in patients with cirrhosis. Provincial, Villarroel 170,08036 Barcelona, Spain. Recently, norfloxacin has been shown to be a uniquely 3111/23062 716

Vol. 12, No. 4, 1990

NORFLOXACIN PREVENTS SPONTANEOUS BACTERIAL PERITONITIS

717

quarter and bimonthly thereafter) in the outpatient clinic. Patients were strongly encouraged to take the study medication as prescribed. They received a new bottle of study medication at each visit and were expected to return the bottle from the previous visit. Compliance was assessed by counting the number of pills returned and by using HPLC (24) to check the presence of norfloxacin in a urine sample obtained at each visit. Patients were defined as noncompliant when (a) the drug had not been taken during five or more consecutive days, (b) 10% or more of the pills were not taken during the entire follow-up and/or (c) norfloxacin was not detected in two or more of the urine samples obtained at each visit. The first of these criteria was chosen considering that significant amounts of norfloxacin are still detectable in feces 5 days after norfloxacin withdrawal (25). The second criteria was arbitrarily chosen. The third criteria was selected as an objective measure to ensure patients would adhere to the treatment. In patients in whom signs of infection developed during followup, a complete physical examination; standard laboratory tests (including complete white blood cell count, fresh urine sedPATIENTS AND METHODS iment and ascitic fluid cell count); chest x-ray film and blood, Patient Population. From December 1987 to August 1989, urine and ascitic fluid cultures were performed. Because the 80 cirrhotic patients who had recovered from an episode of end-point of the trial was SBP recurrence, norfloxacin or SBP were included in a double-blind, multicenter, placebo- placebo administration was definitely interrupted only in controlled study aimed at comparing long-term oral nor- patients who experienceda new episode of SBP. These patients floxacin administration vs. placebo in the prophylaxis of SBP were also excluded from follow-up.Norfloxacin or placebo were recurrence. Diagnosis of SBP was based on the following only transitorily withdrawn in patients suffering from other criteria: (a) abdominal pain and/or fever; (b) ascitic fluid types of infection. In patients in whom other complications polymorphonuclear count > 350/mmY and (c) absence of (ascites, gastrointestinal bleeding or hepatic encephalopathy) clinical, laboratory, radiological or ultrasonographical data developed,the administration of norfloxacin or placebo was not suggesting secondary peritonitis. A positive ascitic fluid intempted. In 12 patients (six from each group) a stool sample was culture was not considered an essential diagnostic criteria because causative organisms are not isolated from ascites in obtained in a sterile container at each visit for quantitative approximately 30% to 50% of patients with well-established analysis of the fecal flora. This analysis was performed SBP (22, 23). Ascitic fluid was cultured by direct innodation according to the method of Wexler and Finegold (26). Briefly, of blood culture bottles, except in 22 patients (27%)in whom 1gm of stool specimen was homogenized in 9 ml of 0.05%yeast it was done by conventional methods because the diagnosis of extract and serially diluted 10-foldin the same solution. From SBP in these cases was made in the hospital’s emergency each dilution, 0.1 ml portions were taken and inoculated in the department where ascitic fluid is not routinely cultured in following media: MacConkey agar (Oxoid Limited, Basingblood culture bottles. SBP was initially treated with cefotaxime stoke, UK), blood agar with nalidixic acid (Oxoid Limited), in all patients. In cases that did not respond to this agent, the Chapman-manitol (Oxoid Limited), Sabouraud’s agar with antibiotic therapy was appropriately changed according to the cloramphenicol (Difco Laboratories, Detroit, MI), Rogosa agar in uitm susceptibility of the isolated bacteria or empirically. (Difco), blood agar (Oxoid Limited), Wilkins-Ghalgren media Resolution of SBP was established when all clinical and (Oxoid Limited) with 5% sheep blood and neomycin (250 laboratory signs of infection disappeared and cultures per- p,g/ml) and kanamycin-vancomycin blood agar. After inocuformed after antibiotic withdrawal were negative. Once SBP lation, the plates were incubated aerobically and anaerobically was cured, and before inclusion into the trial, standard liver (Gaspak jars, BBL Microbiology Systems). The incubation and renal function tests and ascitic fluid cell count and protein time ranged between 1 and 7 days, depending on the media. concentration were measured in all cases. Patients with Aerobic and anaerobic isolates were identified by standard prothrombin time lower than 25%; serum bilirubin greater methods (27). The number of bacteria was expressed as than 10 mg/d or serum creatinine greater than 2 mg/dl; those colony-forming units per gram of feces. The lowest detectable with a history of chronic hepatic encephalopathy or those with colony-forming unit was 2 log/gm of feces. Statistical a d p i e . The calculation of the sample size was clinical, biochemical or echographical data suggesting HCC were not admitted to the study. These criteria were selected to based on the results of the only published study analyzing exclude patients with very short survival expectancy. AU SBP recurrence in cirrhotic patients recovering from an eppatients gave written informed consent to participate in the isode of SBP. In this study, Tit6 et al. (11)showed that the 1-yr study, which was approved by the Investigation and Ethics probability of SBP recurrence in these patients was 69%.Using this figure, we estimated that the sample size needed to detect a Committee of each hospital involved in the trial. Drug Regimen and Follow-upEvaluution. Norfloxacin (400 reduction in SBP recurrence from 70% to 35% at 1yr was 82 mg/day) or placebo administration started 1 wk after the patients (41 patients in each treatment group), accepting an a resolution of SBP. Doses were given orally every 24 hrs. error of 0.05 and a p error of 0.10 in a one-tailed test. The analysis of the results was performed in October 1989. Identical norfloxacin and placebo pills were prepared by Merck Sharp and Dohme Research Laboratories (Hoddeston, UK). The probability of SBP recurrence during follow-up was Each patient’s assigned drug was known only by the pharmacy calculated with the Kaplan-Meier method. Curves were statiss M responsible. After being discharged from the hospital, tically compared using the ManteLCox test. A stratified patients were followed up closely (monthly during the first analysis to adjust for any possible imbalance between the two appropriate drug for long-term SID because i t is incompletely absorbed by the intestine, is highly active against aerobic gram-negative bacilli, has low activity against anaerobic bacteria, rarely causes bacterial resistance and has a low incidence of side effects when administered chronically (18).In a preliminary study i n cirrhotic patients, the oral administration of norfloxacin for 1wk was associated with the disappearance of gram-negative bacilli from the fecal flora without significant changes i n the anaerobic bacteria (19). Finally, oral norfloxacin has proved t o be effective i n preventing gram-negative bacterial infections in patients with acute leukemia and granulocytopenia (20, 21). This article reports the results of a multicenter, double-blind, placebocontrolled trial aimed at investigating whether the long-term administration of norfloxacin prevents SBP recurrence in patients with cirrhosis.

718

GINfiS ET AL.

POLYMORPHONUCLEARS

/

mm3

HEPATOLOGY

TABLE 1. Clinical and laboratory data at the beginning of the study and microbiological characteristics of the index SBP" Variables

Age (yr) Sex (ME) Alcoholic cirrhosis (no.) Serum bilirubin (mg/dl) Serum albumin ( g m L ) Prothrombin time (%) Serum cholesterol (mg/dl) Serum gammaglobulin

50000

5000

Norfloxaciu (n = 40)

Placebo (n = 40)

59 f 1 28/12 23 2.8 f 0.3 29 2 0.9 58 ? 2 118 f 7 24 ? 1

56 f 2 26/14 23 2.8 f 0.3 28 h 0.9 62 f 3 130 ? 7 24 % 1

(&)

500

NORFLOXACIN

PLACEBO

FIG.1. Individual values of polymorphonuclear count in ascitic fluid of the index SBP episode in patients from both groups.

BUN (mg/dl) 25 f 2 Serum creatinine (mg/dl) 1.2 f 0.1 27 Ascitic fluid protein concentration 1 gm/dl (no.) Ascitic fluid polymor14,472 f 2,379 phonuclear count (per mm3) Previous episodes of 1.3 % 0.1 S B P (no.) Organism responsible for the index SBP Gram-negative 23 bacilli Other bacteria 6 Negative culture 11 ________

22 ? 2 1.1 % 0.1 26

17,480 ? 2,836

1.2

f

0.1

23 9 8 _ _ _ _ ~

~

"Dataare expressed as mean f S.E.M.No significantdifference was found in any of the parameters studied. bIncludingthe index SBP.

treatment groups in important prognosticvariables at baseline was performed by means of the Cox regression model (28). These calculations were made using the BMDP statistical package (29). Paired and unpaired Student's t test, and the chi-square test were used for other statistical analysis of the results. Results are presented as mean & S.E.M.

RESULTS Characteristics of the Patients. Of the 80 patients

included in the trial, 40 were assigned to receive norfloxacin and 40 were to receive placebo. The diagnosis of cirrhosis was made by liver biopsy in 60 patients. In the remaining 20 cases, it was based on clinical and laboratory data and exploratory findings (abnormal 99mTc-liver-spleen scan, gastroesophageal varices at endoscopy or signs of chronic liver disease at echography). Most patients had advanced liver disease as indicated by the high frequency of positive history of complications: 53 cases (66%) had had previous episodes of ascites that required hospitalization, 33 cases (41%)of encephalopathy and 32 cases (40%) of gastrointestinal hemorrhage. Sixteen patients (20%) had had one or more episodes of SBP before the index SBP. At inclusion in the study, both groups of patients were similar with respect to age, sex, cause of cirrhosis, liver and renal function tests, ascitic fluid protein and polymorphonuclear concentrations, number of previous episodes of SBP and causative organisms of the index SBP (Table 1 and Figure 1).Four of the 23 alcoholic patients (17%) from the norfloxacin group and 5 of the

23 alcoholic patients (22%) from the placebo group continued to drink during follow-up. Compliance.One patient receiving norfloxacin and two receiving placebo voluntarily abandoned the study within the first month of treatment. A third patient from the placebo group abandoned the study 11 mo after entry because he underwent liver transplantation. These four patients were considered censored at the time they left the study. Seven patients in the norfloxacin group and 7 in the placebo group temporarily stopped taking the medication during 5 or more consecutive days and were considered noncompliant. Causes of noncompliance in the norfloxacin group were: patient's own initiative (three cases), general practitioner's decision (two cases), emergency admission to a hospital not involved in the trial (one case) and intestinal occlusion requiring continuous nasogastric suction (one case). Placebo treatment was stopped because of the patient's initiative in four cases, emergency admission to a hospital not involved in the trial in one and was inadvertently stopped in two cases. The mean number of consecutive days without treatment in these patients was 11.4 2 3.1 in the norfloxacin group and 10.4 & 1.9 in the placebo group (p = NS). The analysis of the two other criteria of compliance used in this investigation did not increase the number of

Vol. 12,No. 4, 1990

719

NORFLOXACIN PREVENTS SPONTANEOUS BACTERIAL PERITONITIS

TABLE 2. Individual changes in the fecal concentration of gram-negative bacilli during long-term administration of nortloxacin and placebo in six patients from each therapeutic group" ~

~

Month

Case

Isolated bacteria

Nofioxacin Escherichia coli

no.

1 2 3 4

5 Klebsiella oxytoca Enterobacter cloaca Citrobmter freundii Hafnia alvei Pseudomonas spp.

Aeromonas hydrophila Placebo Escherichia coli

Klebsiella pneumonia Klebsiella oxytoca Enterobacter cloacae

6 1 2 1 6 3 4 6

2 3 7 9 8 10 11 12 9 11 10 12

Baseline

7.3 9.6 7.2 1.8 8.1 8.4 6.3 4.0 6.0 6.1

1

2

3

6

7

B

11

ND ND ND ND ND ND ND ND ND ND ND ND

ND ND ND ND ND ND ND ND ND ND ND

ND ND ND ND ND

ND ND ND ND ND

ND ND ND ND

ND ND ND ND

ND ND ND

ND ND ND

ND ND ND

ND ND ND

ND ND ND

ND ND ND

3.5h

ND

ND ND

ND ND

ND ND

3.1h

5.4&

ND ND

ND ND

ND ND

6.2

ND ND

7.0 9.0 7.5 9.0

8.3 7.3 8.6

7.4 8.4 7.8

8.0 9.2

5.9

6.3

6.5

ND ND ND ND ND

5.4

ND ND ND

8.2 8.5 9.8 9.8 10.9 9.7 6.0 10.4 6.8 8.5

8.3 9.0 7.4 8.6 10.5 8.3 6.1 8.6 7.8 1.2

8.6 8.0 7.7 7.0 9.8 7.3 6.0 8.5 7.8 7.5

3.2'

ND

ND

7.4

7.4

"Results are expressed as logarithm colony-forming units per gram of feces. ND = not detectable. Missing values correspond to patients who experienced SBP or died during follow-up or who had a follow-up period shorter than 11 mo. *Susceptible to norfloxacin 'Resistant to norfloxacin.

noncompliant patients. Two patients in the norfloxacin group and three in the placebo group failed to take more than 10% of the pills during the entire follow-up period. Moreover, norfloxacin was not detected in two or more of the urine samples obtained at each visit in two additional patients in the norfloxacin group. These seven patients, however, had temporarily stopped their medication during 5 or more consecutive days. Norfloxacin was not detected in the urine of any patient from the placebo group. Eff'a of Treatment on Fecal Flora. The long-term effect of norfloxacin or placebo administration on the fecal flora in six patients from each therapeutic group is illustrated in Tables 2, 3 and 4. Norfloxacin administration was associated with an eradication of aerobic gram-negative bacilli from the fecal flora in all patients and during the entire follow-up. In contrast, no significant changes in the fecal concentration of aerobic gram-negative bacilli were observed in patients receiving placebo. In four patients from the norfloxacin group, gram-negative bacilli not present in fecal baseline cultures were transitorily isolated in follow-up cultures. These organisms were Pseudomonas spp. (three cases)

and Aeromonas hydrophila (one case). Neither norfloxacin nor placebo administration significantly affected the fecal concentration of gram-positive cocci (Table 3). The fecal concentration of anaerobic bacteria and Candida spp. before and during the administration of norfloxacin and placebo are shown in Table 4. Norfloxacin administration was associated with the disappearance or with a clear reduction in the fecal concentration of Clostridium spp. in four of the six cases. However, no changes were observed with respect to other anaerobic bacteria or in relation to Candida spp. The fecal concentration of anaerobic bacteria and Candida spp. remained unchanged in patients receiving placebo. ¤ee of SBP. Fourteen patients from the placebo group (35%)and five from the norfloxacin group (12%) experienced an episode of SBP during follow-up (x" = 5.97;p = 0.014).The mean polymorphonuclear count in these episodes of SBP recurrence was 5,540 1,480/mm3(range = 408 to 17,885)in the placebo group and 19,207 .t 8,590/mm3(range = 410 to 45,000)in the norfloxacin group. Of the 14 SBP recurrences in the placebo group, 10 were caused by aerobic gram-negative

*

720

GINfiS ET AL.

HEPATOLOGY

TABLE 3. Individual changes in the fecal concentration of gram-positivecocci during long-term administration of norfloxacin and placebo in six patients from each therapeutic groupa Month

Isolated bacteria

Norfloxacin Enterococcus faecalis

Streptococcus uiridans

Coagulase-negative staphylococcus

Staphylococcus aureus Placebo Enterococcus faecalis

Streptococcus uiridans

Coagulase-negative staphylococcus Staphylococcus aureus

Caseno.

Baseline

1

2

3

5

7

9

11

8.8 8.4 6.6 8.6

8.5 6.0 6.0 8.7

6.5 9.4 6.5

7.6 7.6 4.0

6.8 5.8 7.8

9.4 8.8 7.1 10.3 9.5

8.3 7.6 6.7 10.5 8.4

7.0 9.4 6.0 10.3

7.3 8.3 6.3 10.0

7.1 6.1 8.1

3.8

ND

4.5

ND

ND

ND

6.7 9.6 8.8 9.6 9.0 6.9 7.0 7.7 9.5 8.7 9.7 ND

ND

7.5 10.2 6.8 8.0 8.0 7.2 10.2 6.3 10.6 7.0 9.8 ND

ND 8.0 5.1 4.7

ND ND ND ND

ND ND 2.8 3.3

ND 4.5

ND 3.7

5.2

3.2

8 9 10 11 12 7 8 9 10 11 12 7

8.4 10.9 9.3 7.5 11.2 9.9 8.8 11.0 7.5 8.6 11.0 3.6

8.8 10.2 8.1 6.7 8.5 8.7 9.0 10.3 9.0 7.5 9.6 ND

9.8 9.0 7.7 6.8 8.5 7.0 9.2 10.2 7.7 7.7 10.7 ND

8.8 9.5 7.2

9

6.5

8.0

5.0

1 2 3 5 6 1 2 3 4 5 6 1

7.8 9.0 6.9 8.3 8.5 9.5 8.5 5.8 9.8 8.5 9.7

5.0

ND

4.3

6.3

6.3

9.4 10.7

8.1 9.9

8.2

7.9 9.1 10.3 6.2

7.3 8.9 10.0

6.1 7.5 9.8

7.9 8.4

ND

2.6

4.8

4.5

ND

ND

8.4

3.0

4.3

"Results are expressed as logarithm colony-forming units per gram of feces. ND = not detectable. Missing values correspond to patients who experienced SBP or died during follow-up or who had a follow-up period shorter than 11 mo.

bacilli (Escherichia coli in eight cases, and Citrobacter freundii and Salmonella spp. in one case each), two by gram-positive cocci (Streptococcus pneumoniae and Streptococcus viridans, respectively),and the remaining two were culture negative. In contrast, only one of the five SBP recurrences in the norfloxacin group was caused by aerobic gram-negative bacilli fE. colil (x" = 8.87; p = 0.0029); the remaining four SBP recurrences observed in this group were caused by gram-positive cocci (Streptococcus sanguis, Streptococcus milleri, S. viridans, and S. pneumoniae, respectively). The SBP recurrence probability curves in patients receiving norfloxacin and placebo are shown in Figure 2 (upper graph). At 1 yr of follow-up, the probability of SBP recurrence was 20% in the norfloxacin group and 68%in the placebo group (p = 0.0063). This difference was caused by the lower rate of SBP recurrence caused by aerobic gram-negative bacilli in patients treated with norfloxacin (3% vs. 60% probability rate at 1 yr of follow-up; p = 0.0013) (Fig. 2, lower graph). In contrast,

no significant difference was observed between both groups with respect to the probability of SBP recurrence caused by other organisms or culture-negative SBP (17% in the norfloxacin group and 20% in the placebo group at 1 yr of follow-up). Three noncompliant patients, two from the placebo group and one from the norfloxacin group, had SBP recurrence during follow-up (E. coli was the causative bacteria in the three cases). When all noncompliant patients (seven from each treatment group) were excluded from the analysis of the results, 1-yrprobability of SBP recurrence was 22% in the norfloxacin group and 68% in the placebo group (p = 0.02), and the 1-yr probability of SBP recurrence caused by aerobic gram-negative bacilli was 0% and 58%, respectively (p = 0.0028). According to the Cox model (28), treatment, serum albumin concentration and prothrombin time were the only variables independently associated with the risk of SBP recurrence. Course of the Disease and Survival. The mean follow-

Vol. 12, No. 4, 1990

721

NORFLOXACIN PREVENTS SPONTANEOUS BACTERIAL PERITONITIS

TABLE 4. Individual changes in the fecal concentration of anaerobic bacteria and Ccrndida epp. during long-term administration of nodosacin and placebo in 6 patients from each therapeutic group" Month Isolated bacteria

Norfloxacin Bacteroides spp.

Baseline

1

2

3

6

9.6 9.6 10.0 9.8 8.1 9.6 7.4 9.7 6.4 8.8 9.3 8.6 6.5 6.7 6.8 6.5 ND 4.9 3.6 5.0 5.5 3.3 6.6 5.4

11.2 9.3 9.1 9.9 9.3 9.4 7.8 8.7 6.2 8.6 8.0 8.0 6.1 ND 6.5 ND ND 6.3 4.0 5.6 6.2 3.0 5.5 4.8

9.8 10.3 9.6 11.2 9.6 9.5 6.0 9.0 5.8 9.0 8.5 7.3 6.2 ND 4.6 ND ND 7.0 4.0 5.4 4.6 4.4 4.7 4.5

11.0 11.2 11.0 10.8 8.7

10.5 8.5 8.8 10.3 11.2

5.5 7.9 7.5 9.1 7.3

9 10

11.3 9.8 10.6 10.6 11.6 12.0 8.1 7.5 8.3 6.8 7.4 9.8 9.4 7.5 6.2 5.5

10.4 9.2 10.4 9.8 10.9 11.6 6.8 8.3 8.0 7.7 7.0 9.0 6.8 6.3 5.6 6.6

11

4.4

12 7 9 10 11 12

8.3 5.3 6.5 3.7 4.0 5.0

9.2 10.8 11.6 10.4 11.3 8.4 7.3 8.5 8.3 8.0 6.8 9.6 8.8 5.5 4.9 6.9 3.9 7.0 5.2 5.1 2.3 4.3 4.2

Case no. 1

2 3 4

Lactobacillus spp

Clostridium spp.

Candida spp.

5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6

7

9

11

9.8 9.3 9.7 9.0

9.7 12.0 9.2 10.0

9.8 9.4 12.3

6.3 6.6 6.7 8.8 8.5

7.4 8.5 7.5 8.7

6.5 8.5 9.2 7.8

6.3 6.3 8.9

5.2 ND ND ND 3.6

7.1 ND ND ND 5.7

7.2 ND ND ND

6.0 2.5

ND ND

5.5 4.9 6.2

4.5 5.5 5.2 4.0 2.5

4.8 4.1 4.8 4.1 3.4

4.5 4.6 4.3 4.0

3.2 4.3 4.5 3.3

3.5 5.8 4.3

11.3 10.7 10.0 9.1

11.0 10.2 10.8

10.6 9.5 10.3

11.5

9.5

6.2 8.0 8.2 6.0

6.9 8.2 7.9

6.7 7.3 8.0

7.3 8.5

7.0

6.3 5.8 5.5 8.0

5.6 3.6 5.0

5.6 4.5 5.5

5.5 7.8

5.8

5.4 4.5 3.0

5.4 4.0

4.2 4.5

3.3

2.0

Placebo Bacteroides spp.

Lactobacillus spp

7 8 9 10 11 12 7 8 9 10 11

Clostridium spp.

12 7 8

Candida spp

4.0

8.1 4.7 4.3 2.8 4.5 3.4

"Results are expressed as logarithm colony-forming units per gram of feces. ND = not detectable. Missing values correspond to patients who experienced SBP or died during follow-up or who had a follow-up period shorter than 11 mo.

up period was significantly higher (p < 0.001) in patients treated with norfloxacin (7.6 ? 1.0 mo) than in those receiving placebo (5.1 + 0.6 mo). This difference was caused by the higher number of patients in the placebo group that were excluded from follow-up because of an SBP recurrence.

Infections other than SBP developed in three patients from the norfloxacin group (7.5%) and five from the placebo group (12%) (p = NS)during follow-up. In the norfloxacin group these infections were a baderemia causedby coagulase-positiveStaphylococcusin a patient with a LeVeen shunt, a bacteremia caused by Pseudo-

722

HEPATOLOGY

GINfiS ET AL.

1

.oj

TABLE 5. Patients readmitted during follow-upand causes of readmiesion" Causes

>

El z;'

m

PLACE60

0.6

2 0.4 0 a n

4

7

II"

p=0.0063

1 .o U F

>

0.8]

=!

0.6

PLACEBO

I

m

a

OL 0

*'{.

Oa21

Jp=O.O013

3

NORFLOXACIN

Ascites SBP recurrenceb Encephalopathy Gastrointestinal hemorrhage Othef

Norftoracin

Placebo

20 (36) 7 (10) 4 (4) 9 (10)

27 (42) 5 (6) 11(11) 6 (10)

3 (3)

8 (8)

8 (9)

6 (7)

"Only the main cause of readmission is considered. The number of readmissions is given in parentheses. patient from the norfloxacin group and three from the placebo group experienced an SBP recurrence during a n admission to hospital for another reason. 'Possible infection (two admissions in one patient) and hemoperitoneum, herpes zoster, cerebrovascular hemorrhage, intestinal occlusion, surgical treatment of an umbilical hernia, diuretic-induced renal failure and pneumonia (one case each) in patients treated with norfloxacin. Possible infection (two patients), severe hyperglycemia (two admissions in one patient), and strangulated hernia, small bowel perforation and digitalis intoxication (one case each) in patients treated with placebo.

MONTHS

lopathy per month of follow-up were 0.17 ? 0.04 and 0.25 0.08, respectively) (t = 0.81, p > 0.10). patients from both groups. Lower graph Cumulative probability of Seven patients from the norfloxacin group and 10 SBP recurrence caused by aerobic gram-negative bacilli in patients from the placebo group died during follow-up. Causes of from both groups. death in these patients were liver failure in eight cases (five from the norfloxacin group and three from the placebo group), gastrointestinal hemorrhage in eight monas cepacea (susceptible in uitro to norfloxacin) in a (two from the norfloxacin group and six from the placebo noncompliant patient 6 days after interrupting nor- group respectively), and small bowel perforation in one floxacin and a community-acquired pneumonia (the patient from the placebo group. Of the 19 patients who causative organism could not be isolated). In the placebo developed SBP recurrence, the infection resolved in 18 group, soft tissue infections developed in two patients, cases after appropriate antibiotic therapy. In the recaused by coagulase-negativeStaphylococcus and Kleb- maining case, the course of SBP recurrence could not be siella pneumoniae, respectively, two urinary tract infec- evaluated because the patient died of a massive gastions caused by E. coli and Enterococcus fmcalis, trointestinal hemorrhage 2 days after admission in the respectively, and one an acute orchitis. In addition, hospital. Four of the patients in whom SBP recurrence possible infections developed in two patients from each resolved, however, died of liver failure during hospitalgroup as defined by fever and/or leukocytosis with ization. Side effects related to noriloxacin were observed in neutrophilia and shift to the left without any other evidence of infection. In one of these latter patients only one patient. An oral and esophageal candidiasis (from the placebo group) S. uiridans was isolated from developed in this patient 7 mo after norfloxacin adminthe ascitic fluid, but the ascitic fluid cell count was istration began, but it was resolved with oral nystatin treatment. normal. Twenty-two patients from the norfloxacin group and DISCUSSION 27 from the placebo group required a total of 36 and 42 The results of this investigation confirm that cirrhotic hospital readmissions, respectively, during follow-up. The number of readmissions per month of follow-up patients who had recovered from SBP are highly were 0.27 2 0.08 and 0.41 & 0.09, respectively (t = predisposed to experience new episodes of this type of 1.13, p > 0.10). Causes of readmission are shown in infection during the course of their disease. In the Table 5. The incidence of hepatic encephalopathy during cirrhotic patients included in the placebo group of this follow-upwas similar in both groups (16 patients treated study, the l-yr probability rate of SBP recurrence was with norfloxacin and 13receiving placebo experienced a 68%, almost identical to the 69% probability rate of SBP total of 31 and 23 episodes, respectively, of hepatic recurrence recently reported by Tit6 et al. (11) in a encephalopathy; the number of episodes of encepha- retrospective investigation including 75 patients. F'IG. 2. Upper graph: Cumulative probability of SBP recurrence in

*

Vol. 12, No. 4, 1990

NORFLOXACIN PREVENTS SPONTANEOUS BACTERIAL PERITONITIS

The efficacy of norfloxacin in producing SID has been constantly demonstrated in studies assessing its effect on the fecal flora in healthy subjects (25, 30-32) and in patients with granulocytopenia (20, 21) or recidivant urinary tract infections (33).In these studies norfloxacin was given during periods ranging between 5 days and 3 mo. These investigations showed that norfloxacin, at doses ranging from 100 to 800 mg/day, produced a marked reduction or disappearance of aerobic gramnegative bacilli from the fecal flora with no significant effects on gram-positive cocci and anaerobic bacteria. Norfloxacin administration in these studies was not associated with overgrowth of resistant, potentially pathogenic bacteria nor with an increase in the fecal concentration of Candida spp. The results of this study in patients with decompensated cirrhosis in whom the effects of norfloxacin (400 mg/day) on the fecal flora were assessed over longer periods of time are identical to those reported by these studies with respect to aerobic gram-negative bacilli, gram-positive cocci and Candidu spp. Norfloxacin administration in this study was associated with a marked reduction or disappearance of Clostridiurn spp. in most patients. This finding is not surprising because the minimal inhibitory concentration of norfloxacin for these organisms is in the range of 2 to 128 pg/ml (341, and the fecal concentration of norfloxacin obtained in cirrhotic patients during treatment with 400 mg/day ranges between 432 and 1,367 p g / p (Gines P. et al, Unpublished data). The concentration of other anaerobic bacteria (Bacteroidesspp. and Lactobacillus spp.), however, was not affected by norfloxacin administration. Anaerobic intestinal flora is thought to be essential to avoid overgrowth of resistant, potentially pathogenic bacteria and Candidu spp (12-14). The preservation of most anaerobic bacteria during long-term norfloxacin administration probably accounts for the low incidence of colonization with resistant aerobic gram-negative bacilli and the lack of effect on the fecal concentration of Candidu spp. observed in this study. In fact, colonization with gramnegative bacilli was only transitorily observed in three of the six patients receiving norfloxacin with follow-up quantification of the fecal flora. In these three cases, gram-negative bacilli were cultured in only six of the 29 specimens obtained and three of these isolates were resistant to norfloxacin. Moreover, in only one of the 40 patients in the norfloxacin group did a fungal infection develop (oral and esophageal candidiasis) and no infections caused by norfloxacin-resistant bacteria were seen. The most important finding of this study is that long-term norfloxacin administration in cirrhotic patients significantly reduces the probability of SBP recurrence. In patients treated with norfloxacin, the l-yr probability of SBP recurrence was 20%,whereas it was 68% in those treated with placebo. This result was exclusively caused by a striking difference in the incidence of SBP recurrence caused by aerobic gramnegative bacilli that occurred in 10 patients from the

723

placebo group and in only one noncompliant patient in the norfloxacin group. The probability of SBP recurrence caused by gram-positive bacteria was identical in the two groups. Two mechanisms may contribute to the efficacy of long-term norfloxacin administration in preventing SBP recurrence caused by aerobic gram-negative bacilli in patients with cirrhosis. The first and most attractive mechanism is the SID caused by this drug (30-33). If, as currently accepted, most SBP in cirrhosis are caused by the passage of enteric bacteria into the general circulation and thereafter into the ascitic fluid (9, lo), it is reasonable to assume that the eradication of gram-negative bacilli from the intestinal flora will be followed by a reduction in the rate of SBP recurrence caused by these organisms. On the other hand, in patients with cirrhosis and ascites, the trough plasma levels of norfloxacin after 7 days of treatment with 400 mg/day (0.51 2 0.10 pg/ml; range = 0.23 to 1.06 pg/ml) (Gin& P. et al, Unpublished data) have been found to be similar or higher than the 90% minimal inhibitory concentration of norfloxacin for most aerobic gramnegative bacilli, which ranges from 0.03 to 3.1 p g / d (18, 34). Therefore the constant presence of bactericidal levels of norfloxacin for gram-negative organisms in serum and, presumably, in ascitic fluid could be an additional mechanism explaining the efficacy of longterm norfloxacin administration in preventing SBP recurrence. The outcome of cirrhotic patients with SBP has markedly improved in recent years. The oldest literature of SBP reported an almost 100% mortality rate during the hospitalization in which SBP was diagnosed (3,3537).This figure was reduced to approximately 50% in a series of patients studied in the late 70’s and early 80’s (5, 23) and to less than 30% in studies published during the past 5 yr (6-8). An early diagnosis and treatment of SBP (5, 7, 9, lo), a more rational use of antibiotics (6)and the improvement in the management of the severe complications occurring in these patients are probably the main reasons that account for this reduction in mortality. In this study, only 5 of the 19 cases (26%)in whom SBP recurrence developed died during the hospitalization in which SBP was diagnosed, thus confirming the improvement in prognosis in patients with SBP. The long-term administration of norfloxacin to cirrhotic patients who had recovered from an episode of SBP may further reduce the contribution of this type of infection in the overall mortality of patients with advanced cirrhosis. In summary, the results of the current study indicate that long-term SID with norfloxacin is a highly effective and safe measure to prevent the recurrence of SBP caused by aerobic gram-negative bacilli in cirrhotic patients. Because SBP recurrence in cirrhotic patients receiving prophylactic norfloxacin is usually caused by gram-positive cocci, the initial antibiotic regimen in these patients should include drugs active against this type of organism. Further studies are necessary to msess

724

GIN)% ET AL.

HEPATOLOGY

elimination of potentially pathogenic bacteria. J Infect Dis 1981;143:644-654. 17. Rimola A, Bory F, Teds J, PBrez-Ayuso RM,Arroyo V, Rod& J. Oral nonabsorbableantibiotics prevent infection in cirrhotics with gastrointestinal hemorrhage. HEPATOLOGY 1985;5:463-467. 18. Wolfson JS, Hooper DC. Norfloxacin: a new targeted fluoroquinolone antimicrobial agent. Ann Intern Med 1988;108:238251. Acknowledgments: We thank Dr. L. Rodrigo, J. Be- 19. Gines P, Rimola A, Marco F, Almela M, Marques JM, Rodamilans M, Jimenez de Ante MT, et al. Oral norfloxacin produces a renguer, A. delVal, E. Lissen, J. Panes, R. Esteban, Mrs. selective bowel decontamination in cirrhotic patients. [Abstract] M. R o m b and S. Roig for their participation in the J Hepatol 1988;7(suppl 1):S136. study, and Mrs. Euliilia Ventura for her secretarial 20. Karp JE, Merz WG, Hendriksen C, Laughon B, Redden T, Bamberger BJ, Bartlett JG, et al. Oral norfloxacin for prevention assistance. We are grateful to Merck, Sharp and Dohme of gram-negative bacterial infections in patients with acute Research Laboratories for supplying us with the norleukemia and granulocytopenia: a randomized, double-blind, floxacin and placebo pills. placebo-controlledtrial. Ann Intern Med 1987;106:1-7. 21. Winston D, Ho W, Champlin RE, Karp J , Bartlett J , Finley R, Joshi J, et al. Norfloxacin for prevention of bacterial infections in REFERENCES granulocytopenicpatients. Am J Med 1987;82:40-46. 1. Rimola A, Bory F, Planas R, Xaubet A. Bruguera M. Rod& J. 22. Runyon BA, Hoefs HC. Culture-negative neutrocytic ascites: a Infecciones bacterianas en la cirrosis hephtica. Gastroenterol variant of spontaneous bacterial peritonitis. H E P A T O L O G Y ~ ~ ~ ~ ; ~ 1209-1211. Hepatol 1981;4:453-458. 2. Pinzello G, Simonetti RG, Craxi A, De Piazza S, Spanb C, Pagliaro 23. Rimola A, Felisart J, Ter6s J , Gatell JM, Jimenez de Anta MT, L. Spontaneous bacterial peritonitis: a prospective investigation in Rodes J. Estudio controladode la eficaciateraphtica de pos pautas predominantly nonalcoholic cirrhotic patients. HEPATOLOGY 1983;3: antibibticas en la peritonitis bacteriana e s p o n h e a de la cirrosis: valor pron6stico de 10s datos bacteriol6gicos. Gastroenterol 545-549. 3. Hoefs JC, Canawati HN, Sapico FL, Hopkins R, Weiner J. Hepatol 1984;7:235-241. 24. Boppana VK,Swanson BN. Determination of norfloxacin, a new Montgomerie JZ. Spontaneous bacterial peritonitis. HEPATOLOGY 1982;2:399-407. nalidixic acid analog, in human serum and urine by high4. Le Carrer M,Poupon RY, Petit J , Ballet F, Darnis J. Les infections performance liquid chromatography. Antimicrob Agents Chemodu liquide ascitique chez le cirrhotique: 6tude clinique et biother 1982;21:808-810. logique de 36 6pisodes observes au cours d'une annee. Gastroen- 25. Pecquet S, Andremont A, Tancrae C. Selective antimicrobial terol Clin Biol 1980;4:640-645. modulation of the intestinal tract by norfloxacin in human 5. Kline MM, McCallum RW, Guth PH. The clinical value of ascitic volunteers and in gnotobiotic mice associated with human faecal fluid culture and leukocyte count studies in alcoholic cirrhosis. flora. Antimicrob Agents Chemother 1986;29:1047-1052. 26. Wexler HM, Finegold SM. Impact of imipenedcilastatin therapy Gastroenterology 1976;70:408-412. 6. Felisart J, Rimola A, Arroyo V, P6rez-Ayuso RM, Quintero E, on normal fecal flora. Am J Med 1985;78(suppl6A):41-46. Gin& P, Rod6s J. Cefotaxime is more effective than is ampicillin- 27. Lenette EH, Balows A, Hausler WJ, Shadomy HJ. Manual of tobramycin in cirrhotics with severe infections. H E P A T O L O G Y ~ ~clinical ~ ~ ; microbiology. 4th ed. Washington, DC: American Society 5~457-462. for Microbiology, 1985:143-461. 7. Runyon BA. Spontaneous bacterial peritonitis: an explosion of 28. Cox DR. Regression models and life tables. J R Stat Soc (B) 1972;34:187-220. 1988;8:171-175. information. HEPATOLOGY 8. Fong TL, Akriviadis EA, Runyon BA, Reynolds TB. Poly- 29. Dixon WJ. BMDP statistical software. Berkeley, CA University of morphonuclear cell count and duration of antibiotic therapy California Press, 1985. in spontaneous bacterial peritonitis. HEPATOLOGY 1989;9:423- 30. De Vries-Hospers HG, Welling GW, van der Waaij D. Norfloxacin 426. for selective decontamination: a study in human volunteers. Prog Clin Biol Res 1985;181:259-262. 9. Rimola A, Arroyo V, Rodes J. Infective complications in acute and chronic liver failure: basis and control. In: Williams R, ed. Liver 31. Edlund C, Bergan T, Josefsson K, Solberg R, Nord CE. Effect of failure. Clinica in critical care medicine. London: Churchill norfloxacin on human oropharyngeal and colonic microflora and Livingstone, 1986:93-111. multiple-dose pharmacokinetics. Scand J Infect Dis 1987;19:11310. Hoefs JC, Runyon BA. Spontaneous bacterial peritonitis. Dis Mon 121. 32. Johnson PC, Ericsson CD, Morgan DR, Dupont HL, Cabada FJ. 1985;31:1-48. 11. Tit6 Ll, Rimola A, Gin& P, Llach J , Arroyo V, Rod& J. Recurrence Lack of emergence of resistant fecal flora during successful of spontaneous bacterial peritonitis in cirrhosis: frequency and prophylaxis of traveler's diarrhea with norfloxacin. Antimicrob predictive factors. HEPATOLOGY 1988;8:27-31. Agents Chemother 1986;3:671-674. 12. Van der Waaij D. Antibiotic choice: the importance of colonization 33. Boerema JBJ, Olthof BT,van Saene HKF. Effects of norfloxacin of the faecal flora in patients with complicated urinary tract resistance. New York Research Studies Press. 13. Van der W d j D. Colonization resistance of the digestive tract: infections. Scand J Infect Dis 1986;4827-31. clinical consequences and implications. J Antimicrob Chemother 34. King A, Warren C, Shannon K, Phillips I. In vitro antibacterial activity of norfloxacin (MK-0366).Antimicrob Agents Chemother 1982;10:263-270. 1982;21:604-607. 14. De Vriea-Hospers HG, Sleijfer DT, Mulder NH, Van der Waaij D, Nieweg HO, Van Saene HKF. Bacteriological aspects of selective 35. Conn HO, Fessel JM. Spontaneous bacterial peritonitis in cirrhosis: variations on a theme. Medicine 1971;50:161-197. intestinal decontamination of the digestive tract as a method of infection prevention in granulocytopenic patients. Antimicrob 36. Correia JP, Corn HO. Spontaneous bacterial peritonitis in cirrhosis: endemic or epidemic?Med Clin North Am 1975;59:963Agents Chemother 1981;19:813-820. 981. 15. Sledfer DT, Mulder NH,de Vries-Hospers HG, Fidler V, Nieweg HO, van der Waaij D, van Saene HKF. Infection prevention in 37. Sbnchez-Tapias JM, Ter6s J , Arroyo V, Bosch J, Bruguera M, granulocytopenic patients by selective decontamination of diRod& J. Infeccibn peritoneal en la cirrosis hepatica con ascitis: cinco GOS gestive tract. Eur J Cancer 1980;16:859-869. de expenencia Gastroenterol Hepatol 1978;1:15-21. 16. Guiot HFL, van der Meer JWM, van Furth R. Selective antimi- 38. Runyon BA. Low-protein concentration ascitic fluid is predisposed crobial modulation of human microbial flora: infection prevention to spontaneous bacterial peritonitis. Gastroenterology 1986;91: in patients with decreased host defense mechanisms by selective 1343-1346.

whether norfloxacin is also effective in preventing SBP in other subsets of patients who are known to be predisposed to this type of infection, such as cirrhotic patients with ascites and low ascitic fluid protein concentration (38) and patients with gastrointestinal hemorrhage (17).

Norfloxacin prevents spontaneous bacterial peritonitis recurrence in cirrhosis: results of a double-blind, placebo-controlled trial.

Eighty cirrhotic patients who had recovered from an episode of spontaneous bacterial peritonitis were included in a multicenter, double-blind trial ai...
958KB Sizes 0 Downloads 0 Views