Journal of Hepatology 1999; 31: 211-283
Printed in Denmark . All rights reserved
Munksgaard . Copenhagen
Development of quinolone-resistantstrains of Escihericliiacoli in stools
of patients with cirrhosis undergoing norfloxacin prophylaxis:
clinical consequences
Jose Ramon Aparicio, Jose Such, Sonia Pascual, Angeles Arroyo’, Joaquin Plazas’, Eva Girona,
Ana Gutierrez, Felix de Vera, Jose Maria Palazbn, Fernando Carnicer and Miguel Perez-Mate0
Liver Unit, Servicio de Medicina Interna, ‘Servicio de Microbiologia, Hospital General Universitario de Alicante, Alicante, Spain
Background/Aim: Norlloxacin prophylaxis decreases
the incidence of bacterial infections in high-risk cir-
rhotic patients, but may promote the development of
quinolone-resistant gram-negative bacteria in stools,
and eventually lead to infections due to these bacteria.
The aim of the study was to evaluate the prevalence
of quinolone-resistant strains of E.coli in stools on ad-
mission, and the characteristics of any nosocomial in-
fections.
Methods: Eighty-three consecutively hospitalized cir-
rhotic patients were included in this prospective study.
The presence of quinolone-resistant strains of E.coli
in stools on admission, and the characteristics of any
nosocomial infections were recorded.
Results: Fourteen out of 83 patients (16.8%) showed
quinolone-resistant E. coli in stools (Group I), and 69
did not (Group II). Thirteen out of 14 from Group I
(92.8%) and 17169 (24.6) from Group II had received
primary or secondary prophylaxis with norfloxacin
(p<O.OOl). During hospitalization, 12/12 (100%) of
patients from Group I and 25/66 (37.8%) of patients
B ACTERIAL infections are a common and potentially
fatal complication of cirrhosis. Thirty-three to
61% of cirrhotic patients may develop bacterial infec-
tions during the natural history of the disease (l-3)
and this may cause death in 25-30% of patients (4).
Spontaneous bacterial peritonitis (SBP) is the most
common infectious complication in cirrhotic patients
Received 28 September 1998; revised 10 February; accepted I5 February
1999
Correspondence: Jest Such, Unidad HepBtica, Servicio de
Medicina Interna, Hospital General Universitario de
Alicante, C/Pintor Baeza s/n, 03010 Alicante, Spain. Tel:
96 593 83 45. Fax: 96 593 83 55. e-mail: jsuchr@coma.es
Copyright 0 European Association
for the Study of the Liver 1999
Journal of Hepatology
ISSNO168-8278
from Group II underwent norfloxacin prophylaxis.
Three bacterial infections in patients from Group I, 3
from Group II patients receiving norlloxacin and 16
from Group II patients not receiving norfloxacin were
recorded (p<O.O5). No infections due to quinolone-
resistant E. coli were observed in patients colonized
with these bacteria. Treatment with norfloxacin in-
duced the development of quinolone-resistant E. coli
in 6/14 (42.8%) patients in a mean time of 18.5~9.8
days.
Conclusions: The development of quinolone-resistant
strains of E. coli is significantly associated with pre-
vious administration of norfloxacin prophylaxis. How-
ever, in our series this fact is not associated with an
increased incidence of quinolone-resistant E. coli or
other gram-negative infections.
Key words: Cirrhosis; Escherichia co& Norfloxacin
prophylaxis; Nosocomial bacterial infections;
Quinolone resistance.
with ascites (1). Most organisms causing bacterial in-
fections in cirrhosis are gram-negative bacilli of intesti-
nal origin, with Escherichia coli being the most com-
monly isolated organism (5).
Cirrhotic patients with low-protein or C3 factor-de-
ficient ascitic fluid (AF) (6,7), those developing epi-
sodes of gastrointestinal bleeding (GIB) (8,9) and those
who have recovered from a previous episode of SBP
(10) are considered to be at high risk of AF infection.
Several studies have shown that norfloxacin prophy-
laxis significantly decreases the incidence of bacterial
infections in these subsets of patients (9-l 1). Two re-
cent metaanalyses have observed an overall decrease of
mortality in patients undergoing antibiotic prophylaxis
277
J. R. Aparicio et al.

fluid infection with a PMN count r-250 per mma, with a positive
(12,13). Moreover, the overall expenses associated with
ascitic fluid culture. Culture-negative neutrocytic ascites (CNNA) was
the development of SBP episodes may be significantly defined as an ascitic fluid infection, with a PMN count ~250 per mmi
reduced in patients receiving norfloxacin prophylaxis with a negative ascitic fluid culture (20). All infections which de-
veloped were recorded and treated according to a predetermined anti-
(14). biogram, or empirically with cefotaxime (1 g/6 h. intravenously) in
Long-term administration of nortloxacin to cir- the case of CNNA. Infections were considered nosocomial when they
rhotic patients. as well to cancer patients with neutro- developed more than 48 h after admission.
penia due to chemotherapy, is associated with the Results are expressed as meantSD. Groups were compared by
means of the unpaired Student’s t-test for continuous variables with
emergence of quinolone-resistant (QR) strains of parametric distribution. the Mann-Whitney U-test for those with
gram-negative bacteria in stools (15,16), and with the non-parametric distribution, and the L’ test for qualitative variables;
applying Yates’ correction when required. Differences were con-
development of bacterial infections due to QR bacteria
sidered significant at the level 0.05. Analyses were performed with the
(17,18). However, little is known about the relationship SPSS Statistical package (SPSS Inc. ver. 6.1.3., 1995, Chicago. IL.
between the presence of QR strains of gram-negative USA).
bacteria in stools and their hypothetical role as patho-
gens in cirrhotic patients. Since E. coli is the most fre- Results
quently isolated gram-negative bacteria causing infec- Fourteen patients were found to show QR strains of E.
tion in these patients, the aims of the present study coli in stools at admission and constituted Group I; 69
were to assess the incidence of QR E. co/i in stools in did not, and constituted Group II.
a cohort of consecutively admitted cirrhotic patients, The clinical characteristics of patients from Groups
and to characterize the spectrum of bacterial infections I and II, reasons for hospitalization and history of pre-
developing during hospitalization in relationship to the vious periods of nortloxacin prophylaxis are detailed
stool QR E. co/i status on admission. in Table 1. No differences were observed in age, sex,
etiology of liver disease, and duration of hospital stay
Materials and Methods between the two groups. Although there was a trend
From March 1996 to March 1997, 83 cirrhotic patients consecutively to an increased severity in the underlying liver disease
admitted to our Liver Unit were included in the study. The diagnosis
of cirrhosis was established by histology or by clinical, laboratory, in patients included in Group I (Child-Pugh C 50%)
endoscopic and/or ultrasonographic findings. The study protocol was compared to Group II (Child-Pugh C 30.4”/;1), this did
approved by the Hospital General Universitario Ethics Comittee, and not reach statistical significance. Similarly, the mean
all patients gave their informed consent to inclusion in the study.
On admission, and prior to any antibiotic treatment, a sample of Child-Pugh score was not statistically different be-
stools was obtained and transported in a Port-a-Cul Universal vial tween the two groups. Ascites and GIB were the main
(Becton Dickinson. Cockeysville, USA). Stools were cultured in 0.1 reasons for hospitalization in both groups. A signifi-
/cg/ml ciprofloxacin-added Agar Mueller-Hinton media (Difco, De-
troit, Michigan, USA). This concentration of ciprolloxacin was cantly higher proportion of admissions due to
chosen according to the recommendations of the Spanish Group encephalopathy was observed in patients from Group
MENSURA (Mesa Espafiola de Normalization de la Susceptibilidad I (28.6% VS. 4.3% respectively, p=O.Ol).
y Resistencia a 10s Antimicrobianos) (19). This concentration can de-
tect with better accuracy those microorganisms with lower levels of Thirteen out of 14 patients included in Group I, and
resistance to quinolones. The cut-off point for resistance to cefotaxi- 17 out of 69 of patients from Group 11 had prophylac-
me has been established in our Institution as I jig/ml.
tically received norfloxacin previously (~~0.001). Thir-
Bacteria growing in ciprofloxacin-added media were identified with
Gram stain and the Vitek System with the GNI card (BioMerieux teen out of 30 patients who had received prior nor-
Vitek, Marcy-I’Etoile, France). The Minimum Inhibitory Conccn- floxacin prophylaxis presented with QR E. coli in
tration (MIC) for ciprofloxacin was determined in those microorgan-
stools at admission (43.3%). QR E. coli was detected
isms identified as E co/i by the E-test (AB Biodisk, Solna. Sweden).
E. coli strains with greater than 0.1 &ml MIC were considered QR. in a similar proportion of patients with previous pri-
Patients were classified according to the presence (Group I) or ab- mary or secondary prophylaxis with norfloxacin [long-
sence (Group II) of QR E. co/i in stools at admission. The proportion
term secondary prophylaxis (519) compared to short-
of patients from Group I and II that had received previous periods
of nortloxacin prophylaxis was recorded. Primary prophylaxis was term primary prophylaxis (S/21), p=NS]. The mean
considered when norfloxacin had been administered due to low-pro- period of norfloxacin administration in patients from
tein ascitic fluid (less than 1 g/d1 total protein) (norlloxacin 400 mgl
both groups receiving secondary prophylaxis was
day during the hospitalization period) or gastrointestinal hemorrhage
(norfloxacin 400 mg/l2 h prr OS or through a nasogastric tube for 7 5.1 t4.6 months.
days after the bleeding episode). Secondary prophylaxis was con- Two patients from Group I (14.3%) and three from
sidered as the continuous administration of norfloxacin to patients
Group II (4.3%) died during the first 24 h after admis-
who had survived a previous episode of SBP (400 mg/day indefinitely
after the recovery from the first SBP episode). sion @=NS), and were not considered thereafter. All
Once admitted, patients were treated according to their clinical sta- patients from Group I and 25 out of 66 (37.8%) from
tus, and Norfloxacin WdS added to the therapeutic protocol when
Group II received prophylactic norfloxacin during the
needed (patients with AF total protein lower than 1 g/dl, and patients
with gastrointestinal hemorrhage). hospital stay. The indications for norfloxacin prophy-
Spontaneous bacterial peritonitis (SBP) was delined as an ascitic laxis during these hospitalizations were similar in both

278
 QR E. coli in stools of patients with cirrhosis

TABLE 1
Clinical characteristics, etiology of liver disease and reasons for admission in patients from Groups I and II
Group I Group II P
No. of patients 14 69
Age (SD); range 61.4 (9.5); 45-79 63.9 (10.7); 32-82 NS
Male/female 1212 48121 NS
Child-Pugh: n (%)
A 3 (21.4) 8 (11.6) NS
B 4 (28.5) 40 (57.9) NS
C 7 (50) 21 (30.4) NS
Child-Pugh mean (SD) 9.1 (2.4) 8.5 (1.9) NS
Reason for admission: n (%)
GIB 6 (42.8) 24 (34.7) NS
Ascites 4 (28.6) 28 (40.5) NS
Encephalopathy 4 (28.6) 3 (4.3) 0.01
SBP 0 2 (2.8) NS
Others 0 12 (17.4) NS
Etiology of liver disease:
Alcohol 10 (71.4) 34 (49.3) NS
HCV 2 (14.3) 20 (29) NS
Unknown 2 (14.3) 8 (11.6) NS
Others 0 7 (9.9) NS
Previous administration of SID
Total 13 (92.8) 17 (24.6) <O.OOl
Primary SID 8 (57.1) 13 (76.5) NS
Secondary 5 (35.7) 4 (23.5) NS
Hospital stay (days?SD) 12.7e7.7 12.3~6.9 NS
SD: standard deviation. NS: not significant. GIB: gastrointestinal bleeding. SBP: spontaneous bacterial peritonitis. HCV hepatitis C virus. SID:
selective intestinal decontamination.

groups. GIB was the main indication for norfloxacin sion: a urinary infection due to E. coli (quinolone and
prophylaxis during hospitalization in both groups (7/ cefotaxime sensitive), and a bacteremia due to K. pneu-
12 and 17/25, respectively), four patients in each group moniue (quinolone and cefotaxime sensitive). These in-
received norfloxacin as secondary prophylaxis of SBP fections were considered community-acquired. Table 3
and five patients (one and four, respectively) had low shows nosocomial bacterial infections recorded in pa-
AF total protein. No differences were observed in the tients from Groups I and II. The incidence of noso-
duration of hospital stay between patients from Group comial bacterial infections was similar in patients in
II who received norfloxacin prophylaxis vs. those who
did not (12.3k7.02 vs. 12.4k7.01 respectively, p=NS).
Table 2 shows the clinical characteristics of patients
TABLE 2
from Group II prophylactically treated or not with
Clinical characteristics of patients from Group II (without quinolone-
norfloxacin during the hospitalization. No differences resistant E. coli at admission) according with the administration of
were observed regarding sex, age and Child-Pugh grade norfloxacin prophylaxis during the hospital stay
of liver insufficiency between the two groups. However, Received Not receiving p
Child-Pugh score was significantly higher in patients norfloxacin norlloxacin
who received norfloxacin prophylaxis during hospital- No. of patients 25 41
ization. Similarly, the median Child-Pugh score of the Age (SD) 64.6 (9.6) 63.5 (11.4) NS
Range 41-79 32-82
overall group of patients receiving prophylasis with 17/S 27114 NS
Male/female
norlloxacin (included in Groups I and II) was signifi-
Child-Pugh: n (%)
cantly higher than that from patients not requiring A 3 (12) 5 (12.2) NS
norfloxacin prophylaxis (9.0752.3 vs. 8.09% 1.5) (p= B 12 (48) 28 (68.3) NS
C 10 (40) 8 (19.5) NS
0.028).
One patient from Group I and one from Group II Child-Pugh score 9.03k2.3 8.092 1.5 0.024
developed an infection during the first 48 h of admis- SD: standard deviation. NS: not significant.

279
J. R. Aparicio et al.

TABLE 3 norfloxacin prophylaxis during the hospitalization de-

Infections developing in patients included in Groups I and II (patients veloped a higher number of infections than patients
with and without quinolone-resistant E. coli at admission respec-
tively)
from Group II receiving prophylactic norfloxacin (39%
vs. 12% respectively, p=O.O3). In our institution only
Group I with Group II with Group II without
norfloxacin norfloxacin norfloxacin patients with lower than I g/d1 AF total protein receive
prophylaxis prophylaxis prophylaxis norfloxacin prophylaxis. However, other studies have
n=12 ?I=25 n=41 established this cut-off at 1.5 g/d1 (11,21). Therefore,
CNNA 2 0 4 patients predisposed to bacterial infections and not re-
Urinary 1 2 5
- ceiving norfloxacin prophylaxis may have been in-
Cutaneous - 2
SBP - 1 2 cluded in this Group. The incidence of infections due
Bacteremia - _ 3 to QR bacteria was similar in patients from Group 11
Total 3” 3b 16
receiving or not norfloxacin prophylaxis (12% vs.
All patients from Group I received norlloxacin prophylaxis. Patients 12.2%, p=NS), and this is probably due to the fact
included in Group II are classified according to whether they received
norfloxacin prophylaxis during their hospital stay. B NS vs infections
that all these infections were caused by gram-positive
developing in Group II with or without norfloxacin. bp=0.03 with bacteria, which are usually resistant to norfloxacin.
respect to infections developing in patients from Group II without Table 4 shows bacteria isolated in infections of pa-
norfloxacin prophylaxis. CNNA: culture-negative neutrocytic ascites.
SBP: spontaneous bacterial peritonitis.
tients from both groups and sensitivity to quinolone
and cefotaxime. All bacteria isolated in infections of
patients submitted to prophylaxis with norfloxacin
TABLE 4
were resistant to quinolones, and 3 out of 4 also were
resistant to cefotaxime. Five out of 11 bacteria causing
Bacteria isolated in patients from Groups I and II
infection in patients not submitted to norfloxacin were
SQ SC Type of infection
QR and two cefotaxime-resistant @=NS in the num-
Group I ber of QR bacteria isolated between patients receiving
Acinetohacter R S Urinary
or not prophylaxis with norfloxacin).
Group II receiving prophylactic norfloxacin
Mortality was similar in patients from both groups
S. aweus R R SBP
E. fecalis R R Urinary (8.3% vs. 9.1% respectively, p=NS). Causes of death
Corynebact spp. R R Urinary were hepatocellular carcinoma in three patients, hepa-
Group II not receiving prophylactic norlloxacin torenal syndrome in one case, GIB in one case, and
E. fecalis R R Urinary liver failure in three patients. No infection-related
E. fticalis R R SBP
S. aweus R S Bacteremia
deaths were observed.
S. aweus R S Urinary Surveillance stool cultures were carried out in 14 pa-
S. aweus R S Urinary tients from Group II (without QR E. coli at admission)
S. epidermidis S S Cutaneous abscess
S. agalactiae s s Cutaneous abscess
while receiving prophylaxis with norfloxacin. During
P. mirahilis s s Urinary hospitalization QR E. coli strains in stools were de-
E. coli s s SBP and urinary tected in 6 out of 14 patients (42.8%) in a mean time
E. coli s s Bacteremia
K. oxytoca s s Bacteremia
of 18.5r9.8 days (range 9-32). Four of these six pa-
tients had received norfloxacin prophylaxis during pre-
All patients from Group I received prophylactic norfloxacin. Group
II patients are subdivided according with the administration of vious admissions to the hospital.
prophylactic norfloxacin. SQ: sensitivity to quinolones. SC: sensitivity
to cefotaxime. S: sensitive. R: resistant,
Discussion
In this prospective study we have evaluated the preva-
lence of QR strains of E. coli on stool cultures of pa-
Group I and Group II (25% W. 28.8% respectively, p= tients with cirrhosis consecutively admitted to a Liver
NS). The incidence of severe infections during hospi- Unit, in a l-year period, the relationship between the
talization, including bacteremia, SBR and CNNA, was presence of QR strains in stools and previous prophy-
also similar in both groups (16.7% vs. 15.2%, respec- lactic treatment with norfloxacin, and the development
tively, p=NS). Although the incidence of infections was of bacterial infections during their hospitalization. To
slightly higher in patients treated with norfloxacin our knowledge, this is the first prospective study of
from Group I than in those from Group II (25% vs. bacterial infections in hospitalized patients with cir-
12%, respectively, p=NS), no differences were observed rhosis in relation with the investigation of the presence
in the number of gram-negative bacilli infections (NS). of QR E. coli in stools at admission.
Interestingly, patients from Group II not submitted to Norfloxacin is an incompletely absorbed fluorinated

280

quinolone with broad-spectrum activity against gram-
negative bacteria. Norfloxacin prophylaxis has proven
to significantly decrease the incidence of infections due
to gram-negative bacteria in neutropenic patients (22),
cirrhotic patients with gastrointestinal hemorrhage (9)
or low-protein ascitic fluid (11,21) and also reduces
SBP recurrence in patients who survive a prior episode
of SBP (10). The decrease in bacterial infections in pa-
tients submitted to norfloxacin prophylaxis is due to a
lower incidence of infections caused by gram-negative
bacteria (23).
Although the development of QR E. coli in stools of
patients receiving norfloxacin or ciprofloxacin as
prophylaxis for bacterial infections has been reported
(24,25), as far as we know no prospective studies have
reported whether the presence of QR strains of E. coli
in stools in patients with cirrhosis at admission has
clinical relevance regarding the type of infections devel-
oping during the hospital stay. Carratala et al. (16)
found that 40% of neutropenic cancer patients receiv-
ing chemotherapy developed QR strains of E. coli in
stools after a median period of norlloxacin administra-
tion of 10 days (range 3 to 35 days). It is of interest to
point out that most (80%) of these patients had re-
ceived fluorquinolones in previous hospital admissions,
but did not present with QR bacteria at admission.
Dupeyron et al. (15) treated 31 cirrhotic patients with
norfloxacin 400 mg QD, and observed that 16 patients
(51.6%) developed QR strains of different gram-nega-
tive bacteria between 12 and 43 days (median 25 days)
of nortloxacin treatment. In contrast, Rolachon et al.
(26) studied the incidence of SBP in a group of 28 cir-
rhotic patients undergoing prophylaxis with ciproflox-
acin 750 mg per OS once a week for 6 months vs. a
placebo group. In this study, no QR resistant strains of
E. coli were detected after 6 months of treatment in a
subgroup of 10 patients from the therapeutic group.
Gin& et al. (10) treated a group of 40 cirrhotic patients
who survived an episode of SBP with norlloxacin 400
mg!day during a mean follow-up of 7 months. No QR
gram-negative strains were found in a series of monthly
stool cultures, and no infections due to QR bacteria
were reported.
In our series, a significantly higher proportion of pa-
tients with QR E. coli in stools at admission had re-
ceived prior prophylaxis with norfloxacin compared
with patients without QR E. coli in stools (92.8% vs.
24.6%, p<O.OOl). From a different point of view,
41.9% of patients who had previously received primary
or secondary norlloxacin prophylaxis presented with
QR E. coli at admission. This figure is similar to that
reported by Carratala et al. (16) and Dupeyron et al.
(15). In addition, 6 out of 14 patients (42.8%) of our
QR E. coli in stools of patients with cirrhosis
study without QR E. coli at admission and treated with
norfloxacin developed QR E. coli in stools in a mean
time of 18.529.8 days (range 9-32), a figure similar to
that reported by other authors (15,16).
As previously reported (9,l l), the incidence of noso-
comial bacterial infections was higher in our patients
not submitted to norfloxacin prophylaxis than in those
prophylactically treated with norfloxacin, despite the
Child-Pugh score being higher in treated patients. Four
out of 37 patients (10.8%) submitted to norlloxacin
prophylaxis and 11 out of 41 patients (26.8%) not re-
ceiving prophylactic norfloxacin during hospitalization
developed culture-proven bacterial infections (p=NS).
No significant differences were observed in the pro-
portion of infections due to gram-positive bacteria be-
tween patients receiving or not receiving nortloxacin
prophylaxis (3/4 and 7/l 1, respectively). The mean
period of hospitalization was similar in patients from
Groups I and II. The fact that in our hospital patients
with cirrhosis are routinely admitted to a Liver Unit
may explain the short mean period of hospitalization
observed in our study.
Although it has been reported (17,18), the devel-
opment of bacterial infections due to QR gram-nega-
tive bacteria does not seem to be a common event.
In a recent work, Tomas et al. showed that the over-
all incidence of bacterial infections in a group of
hospitalized cirrhotic patients undergoing norfloxacin
prophylaxis was 9.6%. These infections were due to
gram-positive cocci in 66.6%, while only one episode
was due to gram-negative bacteria resistant to nor-
floxacin. Moreover, in a recent retrospective study
from Llovet et al. (27), the authors analyzed 229 epi-
sodes of SBP in a 6-year period, and found only 36
bacterial infection episodes in patients undergoing
norlloxacin prophylaxis. Interestingly, none of these
episodes were due to QR strains of gram-negative
bacteria. Novella et al. (18) treated a group of 56 cir-
rhotic patients at high risk of bacterial infections
with continuous administration of norfloxacin 400
mg/day, versus a group of 53 patients that received
norfloxacin prophylaxis only during the hospitaliza-
tion periods. The follow-up was 43~3 weeks, which
is probably the longest reported period of norfloxac-
in prophylaxis in patients with cirrhosis. Ten episodes
of bacterial infection due to E. coli were observed in
the continuous treatment group, of which 9110 (90%)
were QR strains compared to 4/11 (36.3%) QR
strains in the hospitalization-only treated group. The
number of patients developing infections caused by
QR gram-negative bacteria was similar in both
groups (8.9% in group of continuous treatment vs.
7.5% in the hospital treatment patients). None of the
281
J. R. Aparicio et al.

reported infections due to QR E. coli were found to pathogenesis, diagnosis, and management. Gastroenterologist
1995; 3: 311-28.
be resistant to cefotaxime. 6. Runyon BA. Low-protein-concentration ascitic fluid is predis-
In our study, only one bacterial infection due to a posed to spontaneous bacterial peritonitis. Gastroenterology
QR gram-negative and cefotaxime-sensitive bacteria 1986; 91: 1343-6.
7. Such J, Guarner C, Enriquez J, Rodriguez JL, Seres I, Vilardell F.
was observed in a patient with QR E. coli in stools at Low C3 in cirrhotic ascites predisposes to spontaneous bacterial
admission. It should be noted that this did not result peritonitis. J Hepatol 1988; 6: 80-1.
in any apparent harmful consequences with regard to 8. Rimola A, Bory F Teres J, Perez Ayuso RM, Arroyo V, Rodes J.
Oral, nonabsorbable antibiotics prevent infection in cirrhotics
the response to therapy or outcome. with gastrointestinal hemorrhage. Hepatoiogy 1985; 5: 463-7.
The main difference between our study and that of 9. Soriano G, Guarner C, Tomas A, Villanueva C, Torras X, Gon-
Novella et al. (18) is the follow-up. As has been pre- zalez D, et al. Norfloxacin prevents bacterial infection in cir-
rhotics with gastrointestinal hemorrhage. Gastroenterology 1992;
viously reported, bacterial translocation of an organ- 103: 1267-72.
ism is almost always associated with the intestinal over- 10. Gin& P, Rimola A, Planas R, Vargas V Marco F, Almela M, et
growth of that organism in cirrhotic rats with ascites al. Norfloxacin prevents spontaneous bacterial peritonitis recur-
rence in cirrhosis: results of a double-blind, placebo-controlled
(28), but this seems less clear in human patients. It is trial. Hepatology 1990; 12: 71624.
not known if the development of a QR strain of gram- Il. Soriano G. Guarner C, Teixido M, Such J. Barrios J. Enriauez J.
negative bacteria in patients undergoing norfloxacin et al. Selective intestinal decontamination prevents spontaneous
bacterial peritonitis. Gastroenterology 1991; 100: 477-81.
prophylaxis may be followed by its overgrowth in the 12 Bernard B, Grange JD, Nguyen E, Amiot X, Opolon P Poynard
intestinal lumen. T. Antibiotic prophylaxis (AbP) for the prevention of bacterial
From a clinical point of view, the practice of prophy- infections in cirrhotic patients with ascites: a meta-analysis [ab-
stract]. Hepatology 1996; 24: 127lA.
laxis with nortloxacin decreases the incidence of infec-
13 Grange JD, Bernard B, Nguyen E, Opolon P Poynard T Anti-
tions (9,l l), mortality (12,13), and cost (14). Moreover, biotic prophylaxis (AbP) for the prevention of bacterial infections
the rate of resolution and clinical outcome of the infec- in cirrhotic patients with gastrointestinal bleeding (GB): a meta-
analysis [abstract]. Hepatology 1996; 24: 1272A.
tious episodes due to QR bacteria is similar to those
14 Inadomi J. Sonnenberg A. Cost-analysis of prophylactic anti-
caused by quinolone-sensitive bacteria (18,27). There- biotics in spontaneous bacterial peritonitis. Gastroenterology.
fore, although the development of QR strains of gram- 1997; 113: 1289-94.
15. Dupeyron C, Mangeney N, Sedrati L, Campillo B, Fouet P Lelu-
negative bacteria is an undesired side effect of nor-
an G. Rapid emergence of quinolone resistance in cirrhotic pa-
floxacin administration, no harmful clinical conse- tients treated with norfloxacin to prevent spontaneous bacterial
quences have been detected so far. peritonitis. Antimicrob Agents Chemother 1994; 38: 340-4.
In summary, the use of norfloxacin as primary or 16. Carratala J. Fernandez-Sevilla A, Tubau F, Dominguez MA, Gu-
diol E Emergence of fluorquinolone-resistant Escherichia coli in
secondary prophylaxis of bacterial infections in pa- fecal flora of cancer patients receiving norfloxacin prophylaxis.
tients with cirrhosis at high risk for infection signifi- Antimicrob Agents Chemother 1996; 40: 503-5.
cantly increases the prevalence of QR strains of E. coli 17. Castellote J, Xiol J, Rota-Rota R, Fernandez-Esparrach G.
Spontaneous bacterial peritonitis and empyema by Escherichiu
in stools. However, in our series of hospitalized pa- coli resistant to norfloxacine in a patient on selective intestinal
tients with cirrhosis, this is not associated with an in- decontamination with nortloxacin. J Hepatol 1994; 20: 436.
creased incidence of bacterial infections due to QR E. 18. Novella M, Sola R, Soriano G, Andreu M, Gana J. Ortiz J, et
al. Continuous versus inpatient prophylaxis of the first episode
coZi or other gram-negative bacteria in hospitalized cir-
of spontaneous bacterial peritonitis with norfloxacin. Hepatology
rhotic patients, even in the presence of continued nor- 1997; 25: 532-6.
floxacin prophylaxis. 19. Baquero F, Martinez-Beltran J, Canton R. Crieria of the MEN-
SURA group for the definition of the critical points of sensitivity
to antibiotics. Enferm Infect Microbial Clin 1998: 16: 85-92.
Acknowledgement 20. Such J, Runyon BA. Spontaneous bacterial peritonitis. Clin In-
The authors thank Dr. Donald Hillebrand for his in- fect Dis 1998; 27: 669976.
21. Grange D, Roulot D, Pelletier G, Pariente EA, Denis J, Ink 0,
valuable comments.
et al. Norfloxacin primary prophylaxis of bacterial infections in
cirrhotic patients with ascites: a double-blind randomized trial. J
References Hepatol 1998; 29: 430-6.
1.Caly WR, Strauss E. A prospective study of bacterial infections 22. Karp JE, Merz WG, Hendricksen C, Laughon B, Redden T
in patients with cirrhosis. J Hepatol 1993; 18: 353-8. Bamberger BJ, et al. Oral norlloxacin for prevention of gram-
2. Palazon JM, Garcia A, Gomez A. Infecciones hopitalarias en negative bacterial infections in patients with acute leukemia and
pacientes con cirrosis hepatica. Gastroenterol Hepatol 1984; 7: granulocytopenia: a randomized, double-blind, placebo-con-
120-2. trolled trial. Ann Intern Med 1987: 106: l-7.
3. Rimola A. Infecciones bacterianas en la cirrosis hepatica. MTA- 23. Tomas A, Soriano G, Guarner C, Portorreal R, Novella MT, Vil-
Medicina Interna 1987; 5: 161-224. ardell E Hospital-acquired bacterial infections in patients with
4. Felisart J, Rimola A, Arroyo V, Perez Ayuso RM, Quintero E, cirrhosis undergoing selective intestinal decontamination. J
Gines P, et al. Cefotaxime is more effective than is ampicillin- Hepatol 1993; 18: 262-3.
tobramycin in cirrhotics with severe infections. Hepatology 1985: 24. Cometta A, Caiandrd T, Bille J, Glauser MP Escherichia coli
5: 457-62. resistant to fluoroquinolones in patients with cancer and neutrop-
5. Guarner C. Runyon BA. Spontaneous bacterial peritonitis: enia. New Engl J Med 1994: 330: 1240-l.

282

25. Terg R, Llano K, Cobas SM, Brotto C, Barrios A, Levi D, et al.
Effects of oral ciprofloxacin on aerobic gram-negative fecal flora in
patients with cirrhosis: results of short- and long-term administra-
tion, with daily and weekly dosages. J Hepatol 1998; 29: 43742.
26. Rolachon A, Cordier L, Bacq Y, Nousbaum JB, Franza A, Paris
JC, et al. Ciprofloxacin and long-term prevention of spontaneous
bacterial peritonitis: results of a prospective controlled trial. Hep-
atology 1995; 22: 11714.
QR E. coli in stools of patients with cirrhosis
27. Llovet JM, Rodriguez Iglesias P, Moitinho E, Planas R, Bataller
R, Navasa M, et al. Spontaneous bacterial peritonitis in patients
with cirrhosis undergoing selective intestinal decontamination. A
retrospective study of 229 spontaneous bacterial peritonitis epi-
sodes. J Hepatol 1997; 26: 88-95.
28. Guarner C, Runyon BA, Young S, Heck M, Sheikh MY. Intesti-
nal bacterial overgrowth and bacterial translocation in cirrhotic
rats with ascites. J Hepatol 1997; 26: 13728.
283