Introduction
Acute cholangitis is a frequently encountered clinical entity that can have potentially
fatal consequences if not detected and treated promptly. There is a high risk of end-organ
damage, leading to a significant mortality rate if treatment is not instituted promptly
[1], which drops to < 5 % with timely and appropriate interventions [2]. Treatment for acute cholangitis is incumbent on prompt and adequate hemodynamic
resuscitation [3], and initiation of appropriate antimicrobial therapy to limit the local and systemic
inflammatory response to sepsis and to contain the spread of infection [3], followed by the achievement of definitive source control by decompressing the biliary
tree to relieve biliary obstruction [4]. Biliary decompression is most often achieved by techniques applied at endoscopic
retrograde cholangiopancreatography (ERCP). Multiple studies support superior outcomes
and improved mortality with ERCP compared with interventional radiology or surgical
modalities [5]
[6]
[7]
[8]
[9]
[10]
[11]. The selection of appropriate empiric antimicrobial agents in acute cholangitis
is an important cornerstone of medical decision-making that is frequently encountered
in clinical care.
Our group has previously reported the microbiology of bile cultures aspirated at the
time of ERCP in patients with acute cholangitis and cholestasis [12]. In our prior study, the bile aspirates from 160 patients from 1994 to 2000 were
analyzed. In the bile aspirates, Escherichia coli, Enterococcus spp., Streptococcus spp., and Klebsiella pneumoniae were most commonly isolated. Polymicrobial infections and Enterococcus infections were more frequently identified in patients with previously placed plastic
biliary stents. Based on susceptibility data, ciprofloxacin and ceftriaxone were effective
against Gram-negative bacilli in 96 % and 91 % of cases, respectively. This has supported
the use of these antibiotics for routine empiric antibiotic coverage in patients with
suspected acute cholangitis.
With widespread use of antibiotics and the global emergence of multidrug-resistant
organisms, efforts are currently underway to identify microbiological characteristics
and recognize patterns of drug resistance related to intra-abdominal infections to
help in directing prudent antimicrobial choice, early de-escalation to pathogen-directed
therapy, and appropriate termination of therapy. This has prompted prior studies and
new recommendations from medical societies reinforcing these principles [2]
[13]
[14]
[15]
[16].
The primary aim of this study was to describe the microbiology patterns of bile aspirates
obtained at the time of ERCP in patients suspected of having acute cholangitis. Secondary
aims included: (i) the identification of factors that affect the microbiology patterns,
such as the presence/absence of a biliary sphincterotomy, presence/absence of a biliary
stent; and (ii) a description of the presence of microbiological resistance among
the obtained biliary organisms.
Methods
This was a single-center retrospective study. The study was approved by the Institutional
Review Board (IRB) of Indiana University School of Medicine (study #1607577831). Patients
were included in this study if a bile aspirate was collected at the time of ERCP for
clinical suspicion of acute cholangitis, from 1 January 2010 to 31 December 2016. ERCP
procedures were carried out at IU Health University Hospital in Indianapolis, Indiana,
a large tertiary care teaching hospital. An IRB-approved, prospectively maintained
institutional ERCP database and a separate microbiology database were queried to identify
patients. The electronic medical record (Cerner, Kansas City, Missouri, USA) and electronic
procedure reporting system (ProVation, Minneapolis, Minnesota, USA) were reviewed
for the identified patients. Patient demographics, procedure-specific data, and microbial
data were collected for each patient and each biliary aspirate.
At our center, bile aspirates are obtained at the time of ERCP at the discretion of
the endoscopist, most often in the setting of clinically suspected acute cholangitis.
They are routinely drawn for this indication by all endoscopists performing ERCP in
our hospital. Very rarely, if the bile is too thick to aspirate, there are concurrent
data available that may make bile culture redundant (e. g. a blood culture is already
positive for a biliary organism in a patient with pus coming from bile duct), or the
endoscopist forgets to obtain the bile specimen, bile may not be aspirated at the
time of the ERCP.
For this study, clinical acute cholangitis was defined as the presence of a cholestatic
biochemical test profile accompanied by fever, with or without biliary dilatation
on imaging, severe abdominal pain, or leukocytosis. Furthermore, patients noted to
have purulent bile (milky-white discharge exiting the biliary orifice) at the time
of ERCP met the criteria for clinical acute cholangitis. For inpatients undergoing
ERCP for suspected acute cholangitis, blood for two sets of aerobic and anaerobic
blood culture bottles was routinely drawn within 48 hours of admission (at the clinical
discretion of the admitting physician).
Generally, all patients undergoing ERCP for suspected cholangitis have received antibiotics
(either as an inpatient, in the emergency room, or pre-procedurally in the ERCP suite).
If the patient is an inpatient, their inpatient antibiotic regimen is continued periprocedurally
(this is usually intravenous cefepime and metronidazole, or intravenous piperacillin–tazobactam).
If the patient is an outpatient, periprocedural antibiotics are given at the discretion
of the endoscopist and most often include intravenous cefazolin or ciprofloxacin,
depending on the clinical scenario and any medication allergies the patient may have.
The duodenoscope reprocessing techniques during the time period of the study were
consistent, with the exception of a more substantial pre-cleaning step and a change
to double high level disinfection (DHLD) in May 2015 for all duodenoscopes used at
ERCP [17] (Appendix 1s, see online-only Supplementary material). DHLD is one of the four supplementary measures
recommended by the US Food and Drug Administration (FDA) for the reprocessing of reusable
duodenoscopes [18]
[19].
Our techniques for collecting, processing, and reporting biliary aspirates during
ERCP have been previously described [12]. Our techniques for collecting, processing, and reporting blood cultures during
the period of this study are largely unchanged from our previously reported study
[12], except for a few modifications. When blood cultures were obtained, they were obtained
from patients who were admitted as inpatients. Our techniques for collection and processing
of bile cultures and blood cultures are described in Appendix 2s.
Statistical analyses
Data analysis was performed using STATA13 (StataCorp., LLC, College Station, Texas,
USA), SAS9.4 (SAS Inc., Cary, North Carolina, USA), and Excel (Microsoft Corporation,
Redmond, Washington, USA). Descriptive data were reported as median and interquartile
range (IQR) for continuous variables. Student’s t-test was used for the comparison of continuous variables. Categorical variables were
described using frequency and proportion; the 95 %CI of a proportion obtained using
the Clopper–Pearson method was also reported. Chi-squared tests or Fisher’s exact
tests were used for comparisons of categorical data.
A crude odds ratio (OR) and its 95 %CI were calculated to compare the relative odds
of the occurrence of a particular organism in the bile culture, with relation to prior
biliary endoscopic sphincterotomy (ES) and indwelling transpapillary biliary stent.
Hochberg's step-up Bonferroni method was used to adjust for multiple tests performed
for multiple organisms in the bile culture; the adjusted P-values were reported. P-values < 0.05 were considered statistically significant.
Results
During the study period, a total of 20714 ERCPs were performed at our institution,
from which 721 ERCPs with bile aspirates (3.5 %) were identified, from 615 individual
patients with suspected acute cholangitis. Some patients had > 1 ERCP with bile aspiration
during the study period, most often in the setting of primary sclerosing cholangitis.
The median age of patients at the time of ERCP was 65 years (IQR21).
Of the 721 ERCPs with bile aspirates, the cultures of the bile aspirates were positive
in 662 ERCPs (91.8 %), with 403 having been performed on inpatients and 259 on outpatients.
Of the negative bile cultures, 19 procedures were performed on inpatients and 40 on
outpatients. The characteristics of the patients with suspected acute cholangitis
who underwent ERCP with bile aspiration for culture are presented in [Table 1]. Patients who had a positive bile culture were significantly older (median 65 vs.
57 years) and were significantly more likely to have purulent bile or pus exiting
the bile duct at the time of ERCP, as described by the endoscopist (44.7 % vs. 18.6 %).
Table 1
Characteristics of patients with suspected acute cholangitis undergoing endoscopic
retrograde cholangiopancreatography (ERCP) with bile aspiration for culture.
|
Total ERCPs (n = 721)
|
ERCPs with positive bile culture (n = 662)
|
ERCPs with negative bile culture (n = 59)
|
P value
|
|
Age, median (IQR), years
|
65 (21.0)
|
65 (20.0)
|
57 (34.5)
|
< 0.001
|
|
Sex, female, n (%)
|
290 (47.1)
|
263 (39.7)
|
27 (45.8)
|
0.36
|
|
Inpatient, n (%)
|
422 (58.5)
|
403 (60.9)
|
19 (32.2)
|
< 0.001
|
|
Choledocholithiasis, n (%)
|
431 (59.8)
|
402 (60.7)
|
29 (49.2)
|
0.41
|
|
Malignant stricture, n (%)
|
232 (32.2)
|
226 (34.1)
|
6 (10.2)
|
0.002
|
|
Benign stricture, n (%)
|
123 (17)
|
118 (17.8)
|
5 (8.5)
|
0.13
|
|
Primary sclerosing cholangitis, n (%)
|
52 (7.2)
|
49 (7.4)
|
3 (5.1)
|
0.79
|
|
Pus exiting bile duct at time of ERCP, n (%)
|
307 (42.5)
|
296 (44.7)
|
11 (18.6)
|
0.006
|
IQR, interquartile range.
Among the 721 cases in which a bile aspirate was obtained at the time of ERCP, 437
blood cultures were also obtained within 48 hours of presentation, of which 174 (39.8 %)
were positive. Of the cases with positive bile cultures who also had blood drawn for
cultures, the corresponding blood cultures grew bacterial isolates in 40.6 % of instances
(160/394). The microbiological characteristics of the blood cultures obtained are
detailed in [Table 2]. There was no significant difference in the proportion of patients with positive
blood cultures based on whether the corresponding bile culture was positive or negative
at the same encounter (40.5 % [95 %CI35.6 %–45.5 %] vs. 32.6 % [95 %CI19.1 %–48.5 %]).
The most common organism isolated on blood culture was E.coli (49/174; 28.2 %).
Table 2
Microbiological characteristics of blood cultures for patients with positive or negative
bile cultures.
|
Total blood cultures (n = 437)
|
Blood cultures with positive bile culture (n = 395)
|
Blood cultures with negative bile culture (n = 43)
|
P value[1]
|
Adjusted Pvalue[2]
|
|
n
|
n
|
% (95 %CI)[3]
|
n
|
% (95 %CI)[3]
|
|
Positive blood culture
|
174
|
160
|
40.5 (35.6–45.5)
|
14
|
32.6 (19.1–48.5)
|
0.31
|
> 0.99
|
|
Escherichia coli
|
49
|
43
|
10.9 (8.0–14.4)
|
6
|
14.0 (5.3–27.9)
|
0.61
|
> 0.99
|
|
Klebsiella spp.
|
36
|
36
|
9.1 (6.5–12.4)
|
0
|
0 (0–9.2)
|
0.04
|
0.26
|
|
Enterococcus spp.
|
15
|
14
|
3.5 (2.0–5.9)
|
1
|
2.3 (0.1–12.3)
|
> 0.99
|
> 0.99
|
|
Enterobacter spp.
|
10
|
9
|
2.3 (1.0–4.3)
|
1
|
2.3 (0.1–12.3)
|
> 0.99
|
> 0.99
|
|
Gram-negative rods, NOS
|
25
|
21
|
5.3 (3.3–8.0)
|
4
|
9.3 (2.6–22.1)
|
0.29
|
> 0.99
|
|
Other organisms
|
39
|
37
|
9.4 (6.7–12.7)
|
2
|
4.7 (0.6–15.8)
|
0.41
|
> 0.99
|
spp., species; NOS, not otherwise specified.
1
P value from chi-squared test or Fisher’s exact test.
2 Adjusted P value from Hochberg's step-up Bonferroni method.
3 CIs obtained using the Clopper–Pearson method.
Of the 662 positive bile cultures, 81.6 % were polymicrobial. The following pathogens
were identified (as a percentage of total positive bile cultures aspirated; therefore,
owing to polymicrobial cultures, percentages add up to > 100 %): Enterococcus spp. 448 (67.7 %); Klebsiella spp. 295 (44.6 %); E.coli 269 (40.6 %); viridans group streptococci 235 (35.5 %); Candida spp. 189 (28.5 %); Pseudomonas spp. 52 (7.9 %); anaerobes 64 (9.7 %; Clostridium spp. 53 and Bacteroides spp. 11); and Staphylococcus aureus 32 (4.8 %). Select antimicrobial susceptibilities of the Gram-negative organisms
isolated from bile cultures are presented in [Table 3]. Furthermore, SPICE organisms (Serratia spp., Pseudomonas spp., Indole-positive Proteus spp., Citrobacter spp., Enterobacter spp., Providentia spp., Morganella spp., and Acinetobacter spp.) with the potential for the inducible ampC gene (an inducible beta-lactamase-coding gene that is upregulated in response to
ceftriaxone, early generation cephalosporins, and extended-spectrum beta-lactam antibiotics)
were isolated in 251 bile cultures (37.9 %; identified by organism species, not by
ampC genetic testing). Among Enterobacteriaceae isolates, extended-spectrum beta-lactamase
(ESBL)-producing organisms were detected in 7.9 % of isolates and carbapenem-resistant
Enterobacteriaceae (CRE) in 3.6 %. Of S.aureus isolates 50 % were methicillin–resistant (MRSA), and vancomycin resistance (VRE)
was detected in 14.7 % of Enterococcus spp. isolates.
Table 3
Antimicrobial susceptibilities of isolated Gram-negative bacilli from bile cultures.
|
Ampicillin[1]
|
TMP/SMX
|
Ceftriaxone
|
Ciprofloxacin
|
|
Klebsiella pneumoniae
|
|
183/201 (91.0 %)
|
192/201 (95.5 %)
|
176/201 (87.6 %)
|
|
Klebsiella oxytoca
|
|
94/94 (100 %)
|
94/94 (100 %)
|
94/94 (100 %)
|
|
Escherichia coli
|
145/269 (53.9 %)
|
213/269 (79.2 %)
|
238/269 (88.5 %)
|
173/269 (64.3 %)
|
|
Citrobacter freundii complex
|
|
16/17 (94.1 %)
|
15/17 (88.2 %)
|
17/17 (100 %)
|
|
Citrobacter freundii
|
|
11/12 (91.7 %)
|
9/12 (75.0 %)
|
11/12 (91.7 %)
|
|
Other Citrobacter spp.
|
|
15/16 (93.8 %)
|
15/16 (93.8 %)
|
15/16 (93.8 %)
|
|
Pseudomonas aeruginosa
|
|
0/52 (0 %)
|
0/52 (0 %)
|
24/52 (46.2 %)
|
|
Enterobacter cloacae
|
|
52/57 (91.2 %)
|
49/57 (86.0 %)
|
46/57 (80.7 %)
|
|
Enterobacter cloacae complex
|
|
27/28 (96.4 %)
|
25/28 (89.3 %)
|
27/28 (96.4 %)
|
|
Other Enterobacter spp.
|
|
22/23 (95.7 %)
|
22/23 (95.7 %)
|
21/23 (91.3 %)
|
|
Aeromonas spp.
|
|
11/13 (84.6 %)
|
13/13 (100 %)
|
13/13 (100 %)
|
|
Other Gram-negative rods, NOS
|
|
42/48 (87.5 %)
|
45/48 (93.8 %)
|
44/48 (91.7 %)
|
TMP/SMX, trimethoprim sulfamethoxazole; spp., species; NOS, not otherwise specified.
1 Sensitivities to ampicillin were not routinely reported for Klebsiella spp. and other species because of the high prevalence of resistance.
Of the 721 total ERCPs, 459 (63.7 %) had a prior biliary ES and 341 (47.3 %) had a
prior biliary stent. The influence of a prior biliary ES or an indwelling biliary
stent on bile cultures are detailed in [Table 4]. Bile cultures were significantly more likely to be positive when obtained in the
presence of a prior biliary ES (96.1 % [95 %CI93.9 %–97.7 %] vs. 84.4 % [95 %CI79.4 %–88.5 %];
OR4.5). A prior biliary ES significantly increased the probability of VRE in the bile
culture (11.5 % [95 %CI8.8 %–14.8 %] vs. 5.0 % [95 %CI2.7 %–8.3 %]; OR2.5]. Certain
bacteria were more common in the presence of a prior biliary ES: Enterococcus spp. (77.1 % [95 %CI73.0 %–80.9 %] vs. 35.9 % [95 %CI30.1 %–42.0 %]; OR6.0); Enterobacter spp. (19.2 % [95 %CI15.7 %–23.1 %] vs. 7.6 % [95 %CI4.7 %–11.5 %]; OR2.9); Pseudomonas aeruginosa (10.0 % [95 %CI7.4 %–13.1 %] vs. 2.3 % [95 %CI0.8 %–4.9 %]; OR4.8); and
Klebsiella spp. (47.1 % [95 %CI42.4 %–51.7 %] vs. 30.2 % [95 %CI24.7 %–36.1 %]; OR2.1).
Table 4
Associations of prior biliary endoscopic sphincterotomy and indwelling transpapillary
biliary stent with bile culture.
|
Biliary endoscopic sphincterotomy
|
Biliary stent
|
|
Yes
(n = 459)
n (%) [95 %CI]
|
No
(n = 262)
n (%) [95 %CI]
|
Crude odds ratio
(95 %CI)
|
Adjusted Pvalue[1]
|
Yes
(n = 341)
n (%) [95 %CI]
|
No
(n = 380)
n (%) [95 %CI]
|
Crude odds ratio (95 %CI)
|
Adjusted P value[1]
|
|
Positive bile culture
|
441 (96.1) [93.9–97.7]
|
221 (84.4) [79.4–88.5]
|
4.5
(2.6–8.1)
|
< 0.001
|
335 (98.2) [96.2–99.4]
|
327 (86.1) [82.2–89.4]
|
9.0
(3.8–21.3)
|
< 0.001
|
|
Multidrug resistant organisms
|
|
ESBL
|
31 (6.8) [4.6–9.4]
|
14 (5.3) [3.0–8.8]
|
1.3
(0.7–2.5)
|
0.93
|
26 (7.6) [5.0–11.0]
|
19 (5.0) [3.0–7.7]
|
1.6
(0.9–2.9)
|
0.94
|
|
MRSA
|
12 (2.6) [1.4–4.5]
|
4 (1.5) [0.4–3.9]
|
1.7
(0.6–5.4)
|
0.93
|
10 (2.9) [1.4–5.3]
|
6 (1.6)
[0.6–3.4]
|
1.9
(0.7–5.2)
|
0.94
|
|
VRE
|
53 (11.5) [8.8–14.8]
|
13 (5.0) [2.7–8.3]
|
2.5
(1.3–4.7)
|
0.04
|
46 (13.5) [10.0–17.6]
|
20 (5.3) [3.2–8.0]
|
2.8
(1.6–4.9)
|
0.002
|
|
CRE
|
22 (4.8) [3.0–7.2]
|
6 (2.3) [0.8–4.9]
|
2.1
(0.9–5.4)
|
0.93
|
20 (5.9) [3.6–8.9]
|
8 (2.1)
[0.9–4.1]
|
2.9
(1.3–6.7)
|
0.09
|
|
Organisms
|
|
Klebsiella spp.
|
216 (47.1) [42.4–51.7]
|
79 (30.2) [24.7–36.1]
|
2.1
(1.5–2.8)
|
< 0.001
|
182 (53.4) [47.9–58.8]
|
113 (29.7) [25.2–34.6]
|
2.7
(2.0–3.7)
|
< 0.001
|
|
Escherichia coli
|
174 (37.9) [33.5–42.5]
|
95 (36.3) [30.4–42.4]
|
1.1
(0.8–1.5)
|
0.66
|
141 (41.3) [36.1–46.8]
|
128 (33.7) [28.9–38.7]
|
1.4
(1.0–1.9)
|
0.17
|
|
Citrobacter spp.
|
31 (6.8) [4.6–9.4]
|
14 (5.3) [3.0–8.8]
|
1.3
(0.7–2.5)
|
0.66
|
23 (6.7) [4.3–9.9]
|
22 (5.8) [3.7–8.6]
|
1.2
(0.6–2.2)
|
0.94
|
|
Pseudomonas aeruginosa
|
46 (10.0) [7.4–13.1]
|
6 (2.3) [0.8–4.9]
|
4.8
(2.0–11.3)
|
0.001
|
42 (12.3) [9.0–16.3]
|
10 (2.6) [1.3–4.8]
|
5.2
(2.6–10.5)
|
< 0.001
|
|
Enterobacter spp.
|
88 (19.2) [15.7–23.1]
|
20 (7.6) [4.7–11.5]
|
2.9
(1.7–4.8)
|
< 0.001
|
80 (23.5) [19.1–28.3]
|
28 (7.4) [5.0–10.5]
|
3.9
(2.4–6.1)
|
< 0.001
|
|
Enterococcus spp.
|
354 (77.1) [73.0–80.9]
|
94 (35.9) [30.1–42.0]
|
6.0
(4.3–8.4)
|
< 0.001
|
296 (86.8) [82.7–90.2]
|
152 (40.0) [35.0–45.1]
|
9.9
(6.8–14.4)
|
< 0.001
|
|
Aeromonas spp.
|
10 (2.2) [1.0–4.0]
|
3 (1.1) [0.2–3.3]
|
1.9
(0.5–7.0)
|
0.66
|
10 (2.9) [1.4–5.3]
|
3 (0.8)
[0.2–2.3]
|
3.8
(1.0–13.9)
|
0.17
|
|
Other Gram-negative rods, NOS
|
47 (10.2) [7.6–13.4]
|
11 (4.2) [2.1–7.4]
|
2.6
(1.3–5.1)
|
0.03
|
42 (12.3) [9.0–16.3]
|
16 (4.2) [2.4–6.7]
|
3.2
(1.8–5.8)
|
0.001
|
|
Anaerobes
|
35 (7.6) [5.4–10.4]
|
29 (11.1) [7.5–15.5]
|
0.7
(0.4–1.1)
|
0.66
|
30 (8.8) [6.0–12.3]
|
34 (8.9) [6.3–12.3]
|
1.0
(0.6–1.6)
|
0.94
|
ESBL, extended spectrum beta-lactamase-producing Enterobacteriaceae; MRSA, methicillin-resistant
Staphylococcus aureus; VRE, vancomycin-resistant Enterococcus; CRE, carbapenem-resistant Enterobacteriaceae; ssp., species; NOS, not otherwise
specified.
1 Adjusted Pvalue from Hochberg's step-up Bonferroni method.
Likewise, the presence of a biliary stent significantly increased the probability
of VRE (13.5 % [95 %CI10.0 %–17.6 %] vs. 5.3 % [95 %CI3.2 %–8.0 %]; OR2.8). A similar
effect regarding the increased presence of certain organisms was seen in the presence
of a prior biliary stent: Enterococcus spp. 86.8 % (95 %CI82.7 %–90.2 %) vs. 40.0 % (95 %CI35.0 %–45.1 %; OR9.9); Enterobacter spp. 23.5 % (95 %CI19.1 %–28.3 %) vs. 7.4 % (95 %CI5.0 %–10.5 %; OR3.9); P.aeruginosa 12.3 % (95 %CI9.0 %–16.3 %) vs. 2.6 % (95 %CI1.3 %–4.8 %; OR5.2); Klebsiella spp. 53.4 % (95 %CI47.9 %–58.8 %) vs. 29.7 % (95 %CI25.2 %–34.6 %; OR2.7). There was
no effect on the incidence of E.coli with regard to the presence of a biliary stent or anaerobes with regard to the presence
of biliary ES or a biliary stent.
Discussion
This single-center retrospective study evaluated the microbiology of biliary aspirates
obtained from 721 ERCP procedures in patients with suspected acute cholangitis. Positive
bile cultures were found in nearly 92 % of these ERCPs. In general, duodenoscope cleaning
techniques, culture appropriation techniques, and processing have remained predominantly
consistent, with only minor variations [17] from the time of a prior study on this topic published by our institution in 2002
[12].
This study covered a 7-year time period between 2010 and 2016, similar to the previously
published bile culture data from our institution for a comparable group of patients
covering a 7-year time period between 1994 and 2000. The current study however contains
over four times more ERCPs with bile aspiration performed than in the prior study
(721 ERCP procedures compared with 180 ERCP procedures), owing to near universal adoption
of the practice by all gastroenterologists performing ERCP for this indication at
our institution.
The current study confirms and expands upon the prior data that there is an increased
incidence in polymicrobial bile cultures, with an increased likelihood of Enterococcus spp. in the setting of an indwelling biliary stent [12]. We also found a significantly higher likelihood of Enterobacter spp. (23.5 % vs. 7.4 %), P.aeruginosa (12.3 % vs. 2.6 %), and K.pneumoniae (53.4 % vs. 29.7 %) in the presence of a biliary stent, which was not appreciated
in the prior study. We suspect this association was uncovered in this study because
of the much larger sample size and expanded culturing of bile in nearly all patients
suspected of having acute cholangitis.
The current study also realized a higher positive bile culture rate in the patients
who had not had a prior biliary stent. We found that bile cultures were positive in
86.1 % of patients who had previously not had a biliary stent, whereas only 55 % of
patients in the prior study had a positive bile culture without a prior biliary stent.
We suspect this may be related to more patients undergoing ERCP with biliary aspiration
who have possibly had other sources of their clinical symptoms, longer duration of
antibiotics pre-ERCP, or different antibiotics pre-ERCP, or to there being less resistance
to antibiotics in biliary organisms in the prior study.
In the current study, we sought to clarify the influence of a prior biliary ES on
the positivity rate and microbiology of bile cultures. We found effects of biliary
ES that were similar to that of having a prior indwelling biliary stent. This is to
be expected as both biliary ES and an indwelling biliary stent violate the natural
sphincter of Oddi barrier between the duodenal lumen and the biliary system [20]. We hypothesize the changes in microbiology with regard to the presence or absence
of a biliary ES or a biliary stent is likely due to migration of luminal bacteria
into the biliary tree and/or colonization of the existing biliary stent. This is further
supported by the increased presence, in these subsets of patients, of organisms that
are well-known to produce biofilms and colonize foreign bodies such has urinary catheters
and biliary stents (Enterobacter spp., Enterococcus spp., Klebsiella spp., and Pseudomonas spp.) [21]
[22]
[23]
[24]. There is a large overlap of patients with a biliary ES and an existing or prior
biliary stent. In addition, although the data are not completely available to analyze
within this study, patients who receive a biliary ES or biliary stent likely have
a higher disease burden from more complex disease, prior hospitalizations, and multiple
exposures to antibiotics, which may contribute to higher rates of positive bile culture,
antibiotic resistance, and bacterial colonization.
We discovered a notable number of high concern multidrug-resistant organisms (155/662;
23.4 %). There was a significant increase in isolation of VRE in the presence of a
biliary ES (11.5 % vs. 5 %; OR2.5 [95 %CI1.3–4.7]; adjusted P = 0.04] and biliary stents [13.5 % vs. 5.3 %; OR2.8 [95 %CI1.6–4.9]; adjusted P = 0.002], but not other multidrug-resistant organisms. This association has been
reported previously [25].
A potential benefit of collecting and culturing bile from patients with suspected
cholangitis is the ability to tailor antibiotic regimens to the resistance pattern
of the biliary aspirates. Furthermore, in patients with recurrent cholangitis from
structural disease (e. g. primary sclerosing cholangitis), the presence of recurrent
positive biliary cultures or highly resistant biliary cultures may prompt liver transplant
teams to advocate for exception points for liver transplantation. We also believe
that, in the era of known risk of transmission of organisms from reusable duodenoscopes,
the knowledge of biliary culture data can inform the decision regarding enhanced endoscope
reprocessing and/or future endoscope usage for that patient. In a patient with a known
multidrug-resistant organism on bile culture, we would consider reprocessing that
endoscope subsequently with an enhanced treatment (ethylene oxide sterilization).
Furthermore, we would consider future ERCPs for that patient being done with a single-use
duodenoscope.
The recommended choice of empiric antibiotics by medical societies for community acquired
biliary infections include a third-generation cephalosporin, a fluoroquinolone, or
a penicillin/beta-lactamase inhibitor [2]
[13]
[26]. In our earlier study, given the 96 % susceptibility of Gram-negative organisms
to ciprofloxacin, we continued to consider ciprofloxacin a recommended initial empiric
choice for suspected acute cholangitis. In the current study however the susceptibility
to ciprofloxacin of the most common Gram-negative organisms has been reduced substantially
(E.coli susceptibility 64 % and K.pneumoniae susceptibility 88 %). This decrease in susceptibility to ciprofloxacin demonstrates
an increase in the prevalence of Gram-negative organisms resistant to this commonly
used antibiotic in our community. A similar change in resistance pattern has been
reported in other geographic areas as well [27]
[28]
[29]. An important strength of this study is that we were able to assess a longitudinal
change over time in the antibiotic sensitivity of biliary organisms by comparing with
a similar cohort from our prior study [12].
Antibiotic selection for suspected cholangitis is an important decision made routinely
in clinical practice. We support a judicious use of empiric antimicrobial agents for
biliary infections that should be modified based on the local antibiogram. We no longer
use ciprofloxacin as an empiric antibiotic for suspected acute cholangitis. For inpatients
suspected of having acute cholangitis, the practice at our institution is to recommend
the empiric use of intravenous fourth-generation cephalosporin (i. e. cefepime) and
intravenous metronidazole, as the combination is active against the most commonly
isolated biliary organisms and also against the SPICE organisms (Serratia spp., Pseudomonas spp., Indole-positive Proteus spp., Citrobacter spp., Enterobacter spp., Providentia spp., Morganella spp., and Acinetobacter spp.), which may have inducible resistance (amp-C gene) to intravenous extended-spectrum beta-lactam antibiotics, such as piperacillin–tazobactam
and ceftriaxone. After ERCP with effective biliary drainage, a decision is made to
de-escalate antibiotics to narrower spectrum oral options, pending evaluation of clinical
improvement and preliminary bile culture and blood culture data. These preliminary
culture data are available in 2–3days and finalized within 5days. As a biliary infection
with isolated Enterococcus spp. is unusual in the presence of an intact sphincter of Oddi barrier between the
duodenal lumen and the biliary system, and polymicrobial infections with Enterococcus spp. do not require enterococcal coverage [26], we do not recommend routine empiric or directed antibiotic coverage for this organism.
The limitations of this study include that it is a single-center retrospective study
conducted at a tertiary academic healthcare center. Given the retrospective nature
of the study, it is possible that a small percentage of patients undergoing ERCP for
suspected cholangitis did not have bile aspirated at the time of ERCP, which may introduce
selection bias. Although we do not have data on the frequency of this occurring, we
believe it is quite rare, given our clinical practice patterns. As microbial frequency
and antimicrobial resistance patterns do have some geographic and regional variation,
this may limit the generalizability of this study. Because all of the patients studied
were at a tertiary academic healthcare center, the findings may not be applicable
to community healthcare centers. However, as our patient population, and procedural
and processing techniques have remained similar since a prior study performed on this
topic at our institution, a historical control is available for this study, which
allows a comparison of data over time.
As most patients are transferred from referring hospitals, we could not quantify all
prior antibiotic exposure in this study or assess whether antibiotics were given prior
to collection of blood or bile cultures. Antibiotic exposure during hospitalization
prior to ERCP likely impacts the positivity of both bile and blood cultures, as well
as the microbiology profile. For instance, bile cultures were negative in 11 of the
307 patients (3.6 %) with pus visualized. This could be due to patients receiving
antibiotics prior to the ERCP, reducing the positivity of the cultures. Another factor
could be a descriptive error by the endoscopist in describing non-purulent dark or
murky bile as pus.
In summary, this study demonstrates the susceptibilities of E.coli and Klebsiella spp. isolates to ciprofloxacin are significantly lower than historically noted. There
was a substantial presence of serious antimicrobial-resistant organisms (ESBL-producing
Enterobacteriaceae, MRSA, VRE, and CRE) identified in this study. The presence of
an endobiliary stent or biliary ES is associated with differences in the microbiological
characteristics and antimicrobial resistance patterns of the isolates. This is the
only study to our knowledge where comparable cohorts of patients have been studied
over time to demonstrate the changes in microbial frequency and changes to antimicrobial
resistance patterns.
To develop an antibiogram, promote prompt titration of antimicrobial therapy, and
decrease the risk of development of antimicrobial resistance, we advocate a biliary
aspirate at the time of endobiliary interventions for suspected acute cholangitis
and suggest it may be a marker for quality improvement at endoscopy centers worldwide.
Furthermore, with recent appreciation of the risk of transmission of drug-resistant
organisms via reusable duodenoscopes, we have used bile culture data to play a part
in our algorithm for the reprocessing of reusable scopes and the selection of single-use
scopes in selected patients.
